rickub
bots-garden/mini-mepublic Fork
⑂ main
Branches
feature/agents-file feature/ask-agent-hooks feature/ask-cli feature/decider-no-tools feature/declared-tools feature/display-called-hooks feature/fetch-build-in feature/genkit-v1.13.1 feature/github-mcp-in-sbx feature/go-1.27.3 feature/hooks feature/improve-init feature/interactive-prompt feature/last-question feature/llmman-provider feature/log-errors feature/lsp-builtin feature/mcp-gateway-session feature/mcp-support feature/memory-search feature/mini-me-as-a-sdk feature/mini-mes feature/mm-init feature/open-terminal feature/prompt-flag feature/response-request feature/resume-acp feature/sbx-template-kit feature/search-in-doc feature/selectors feature/skill-discovery feature/structured-output feature/system-one feature/talk-to-another-mm feature/tasks-builtin feature/try-again feature/unload-model feature/wasm-custom-tools feature/yolo feature/zed-skill-selector fix/generation-timeout fix/plan-mode fix/too-long-completion fix/tool-order fix/word-wrap improve-mem-usage main thinking
Tags
v0.0.0 v0.0.1 v0.0.2 v0.0.3 v0.0.4 v0.1.0 v0.1.1 v0.1.2 v0.1.3 v0.1.4 v0.1.5 v0.1.6 v0.1.7 v0.1.8 v0.1.9 v0.2.0 v0.2.1 v0.2.2 v0.2.3 v0.2.4 v0.2.5 v0.2.6 v0.3.0 v0.4.0 v0.5.0 v0.6.0 v0.7.0 v0.8.0 v0.9.0 v1.0.0 v1.0.1 v1.0.2 v1.0.3 View all branches → View all tags →
Commits
⬇ Clone ▾
git clone https://git.rickub.com/bots-garden/mini-me.git
git clone ssh://git@rickub.com/bots-garden/mini-me.git

Host key fingerprint (ed25519): SHA256:iycHnxEyq0Q7uyVpB7JlznP0G7JrTPXLYRcAU5CSLhc — verify it before your first connect.

Download ZIP

yolo mode on/off #23

Merged@k33g wants to merge feature/yolo into main
Merged as 58ba549b2dc1.
118 files changed
modified .tickets/epics.yaml +6 -0
@@ -12,3 +12,9 @@ epics:
1212 - title: Ideas
1313 state: open
1414 color: "#bfdadc"
15+ - title: Agent Memory
16+ state: open
17+ color: "#f9d0c4"
18+ - title: Providers (LLM Engine)
19+ state: open
20+ color: "#0e8a16"
@@ -12,3 +12,9 @@ epics:
12 - title: Ideas12 - title: Ideas
13 state: open13 state: open
14 color: "#bfdadc"14 color: "#bfdadc"
15+ - title: Agent Memory
16+ state: open
17+ color: "#f9d0c4"
18+ - title: Providers (LLM Engine)
19+ state: open
20+ color: "#0e8a16"
modified .tickets/issues/0013-activate-deactivate-yolo-mode.yaml +15 -3
@@ -1,13 +1,25 @@
11 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
22 id: 13
33 title: Activate / Deactivate YOLO mode
4-state: open
4+state: closed
55 author:
66 name: k33g
77 email: ph.charriere@gmail.com
88 createdAt: 2026-09-16T05:47:28.194Z
9-updatedAt: 2026-10-06T02:52:48.473Z
9+updatedAt: 2026-10-06T03:53:37.224Z
1010 labels: []
11-body: ""
11+related:
12+ - 59
13+body: |
14+ User's request (2026-10-06, FR): enable/disable yolo mode for tool
15+ execution through the config file. Default: yolo. With yolo off, every
16+ tool call asks: once, never, or for the whole session. Examples:
17+ `yolo: false` + `bashTool: true` (asks), `yolo: true` + `bashTool: true`,
18+ `yolo: true` + `bashTool: false` (bash disabled, other tools not
19+ necessarily). With `mmConfigs`, the main agent (the one holding
20+ mmConfigs) wins.
21+
22+ Implemented on branch `feature/yolo` (2026-10-06) — details in ticket
23+ 59's comment 2.
1224 tasks: []
1325 comments: []
@@ -1,13 +1,25 @@
1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
2 id: 132 id: 13
3 title: Activate / Deactivate YOLO mode3 title: Activate / Deactivate YOLO mode
4-state: open4+state: closed
5 author:5 author:
6 name: k33g6 name: k33g
7 email: ph.charriere@gmail.com7 email: ph.charriere@gmail.com
8 createdAt: 2026-09-16T05:47:28.194Z8 createdAt: 2026-09-16T05:47:28.194Z
9-updatedAt: 2026-10-06T02:52:48.473Z9+updatedAt: 2026-10-06T03:53:37.224Z
10 labels: []10 labels: []
11-body: ""11+related:
12+ - 59
13+body: |
14+ User's request (2026-10-06, FR): enable/disable yolo mode for tool
15+ execution through the config file. Default: yolo. With yolo off, every
16+ tool call asks: once, never, or for the whole session. Examples:
17+ `yolo: false` + `bashTool: true` (asks), `yolo: true` + `bashTool: true`,
18+ `yolo: true` + `bashTool: false` (bash disabled, other tools not
19+ necessarily). With `mmConfigs`, the main agent (the one holding
20+ mmConfigs) wins.
21+
22+ Implemented on branch `feature/yolo` (2026-10-06) — details in ticket
23+ 59's comment 2.
12 tasks: []24 tasks: []
13 comments: []25 comments: []
modified .tickets/issues/0030-how-to-use-mini-me-with-claude.yaml +2 -1
@@ -6,8 +6,9 @@ author:
66 name: k33g
77 email: ph.charriere@gmail.com
88 createdAt: 2026-09-21T04:00:56.971Z
9-updatedAt: 2026-09-22T20:57:22.310Z
9+updatedAt: 2026-10-06T04:00:56.602Z
1010 labels: []
11+epic: Providers (LLM Engine)
1112 body: |
1213 ## Analysis (Claude, 2026-09-21 — not implemented)
1314
@@ -6,8 +6,9 @@ author:
6 name: k33g6 name: k33g
7 email: ph.charriere@gmail.com7 email: ph.charriere@gmail.com
8 createdAt: 2026-09-21T04:00:56.971Z8 createdAt: 2026-09-21T04:00:56.971Z
9-updatedAt: 2026-09-22T20:57:22.310Z9+updatedAt: 2026-10-06T04:00:56.602Z
10 labels: []10 labels: []
11+epic: Providers (LLM Engine)
11 body: |12 body: |
12 ## Analysis (Claude, 2026-09-21 — not implemented)13 ## Analysis (Claude, 2026-09-21 — not implemented)
13 14
modified .tickets/issues/0031-how-to-use-mini-me-with-gemini.yaml +2 -1
@@ -6,8 +6,9 @@ author:
66 name: k33g
77 email: ph.charriere@gmail.com
88 createdAt: 2026-09-21T04:01:15.166Z
9-updatedAt: 2026-09-22T20:56:55.345Z
9+updatedAt: 2026-10-06T04:01:18.848Z
1010 labels: []
11+epic: Providers (LLM Engine)
1112 body: |
1213 ## Analysis (Claude, 2026-09-21 — not implemented)
1314
@@ -6,8 +6,9 @@ author:
6 name: k33g6 name: k33g
7 email: ph.charriere@gmail.com7 email: ph.charriere@gmail.com
8 createdAt: 2026-09-21T04:01:15.166Z8 createdAt: 2026-09-21T04:01:15.166Z
9-updatedAt: 2026-09-22T20:56:55.345Z9+updatedAt: 2026-10-06T04:01:18.848Z
10 labels: []10 labels: []
11+epic: Providers (LLM Engine)
11 body: |12 body: |
12 ## Analysis (Claude, 2026-09-21 — not implemented)13 ## Analysis (Claude, 2026-09-21 — not implemented)
13 14
modified .tickets/issues/0033-invoke-another-mini-me.yaml +2 -2
@@ -1,12 +1,12 @@
11 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
22 id: 33
33 title: Invoke (another) Mini Me
4-state: open
4+state: closed
55 author:
66 name: k33g
77 email: ph.charriere@gmail.com
88 createdAt: 2026-09-21T04:48:03.596Z
9-updatedAt: 2026-09-29T06:25:43.079Z
9+updatedAt: 2026-10-06T03:55:11.169Z
1010 labels: []
1111 milestone: v0.3.0
1212 epic: Builtin tools
@@ -1,12 +1,12 @@
1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
2 id: 332 id: 33
3 title: Invoke (another) Mini Me3 title: Invoke (another) Mini Me
4-state: open4+state: closed
5 author:5 author:
6 name: k33g6 name: k33g
7 email: ph.charriere@gmail.com7 email: ph.charriere@gmail.com
8 createdAt: 2026-09-21T04:48:03.596Z8 createdAt: 2026-09-21T04:48:03.596Z
9-updatedAt: 2026-09-29T06:25:43.079Z9+updatedAt: 2026-10-06T03:55:11.169Z
10 labels: []10 labels: []
11 milestone: v0.3.011 milestone: v0.3.0
12 epic: Builtin tools12 epic: Builtin tools
modified .tickets/issues/0034-anatolian-mini-me-s-orchestrator.yaml +2 -2
@@ -1,12 +1,12 @@
11 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
22 id: 34
33 title: Anatolian -> Mini Me(s) Orchestrator?
4-state: open
4+state: closed
55 author:
66 name: k33g
77 email: ph.charriere@gmail.com
88 createdAt: 2026-09-21T04:49:57.129Z
9-updatedAt: 2026-09-21T04:49:57.129Z
9+updatedAt: 2026-10-06T03:57:06.058Z
1010 labels: []
1111 epic: Ideas
1212 body: ""
@@ -1,12 +1,12 @@
1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
2 id: 342 id: 34
3 title: Anatolian -> Mini Me(s) Orchestrator?3 title: Anatolian -> Mini Me(s) Orchestrator?
4-state: open4+state: closed
5 author:5 author:
6 name: k33g6 name: k33g
7 email: ph.charriere@gmail.com7 email: ph.charriere@gmail.com
8 createdAt: 2026-09-21T04:49:57.129Z8 createdAt: 2026-09-21T04:49:57.129Z
9-updatedAt: 2026-09-21T04:49:57.129Z9+updatedAt: 2026-10-06T03:57:06.058Z
10 labels: []10 labels: []
11 epic: Ideas11 epic: Ideas
12 body: ""12 body: ""
modified .tickets/issues/0040-better-list-of-detected-tools-at-the-start-of-mi.yaml +2 -1
@@ -6,8 +6,9 @@ author:
66 name: k33g
77 email: ph.charriere@gmail.com
88 createdAt: 2026-09-23T10:28:36.392Z
9-updatedAt: 2026-09-23T10:28:36.392Z
9+updatedAt: 2026-10-06T04:01:48.816Z
1010 labels: []
11+epic: Improve TUI
1112 body: ""
1213 tasks: []
1314 comments: []
@@ -6,8 +6,9 @@ author:
6 name: k33g6 name: k33g
7 email: ph.charriere@gmail.com7 email: ph.charriere@gmail.com
8 createdAt: 2026-09-23T10:28:36.392Z8 createdAt: 2026-09-23T10:28:36.392Z
9-updatedAt: 2026-09-23T10:28:36.392Z9+updatedAt: 2026-10-06T04:01:48.816Z
10 labels: []10 labels: []
11+epic: Improve TUI
11 body: ""12 body: ""
12 tasks: []13 tasks: []
13 comments: []14 comments: []
modified .tickets/issues/0048-add-a-builtin-tool-to-search-into-the-mm-session.yaml +2 -1
@@ -6,9 +6,10 @@ author:
66 name: k33g
77 email: ph.charriere@gmail.com
88 createdAt: 2026-09-27T03:45:48.121Z
9-updatedAt: 2026-09-29T06:25:13.129Z
9+updatedAt: 2026-10-06T03:59:41.259Z
1010 labels: []
1111 milestone: v0.3.0
12+epic: Agent Memory
1213 body: ""
1314 tasks: []
1415 comments: []
@@ -6,9 +6,10 @@ author:
6 name: k33g6 name: k33g
7 email: ph.charriere@gmail.com7 email: ph.charriere@gmail.com
8 createdAt: 2026-09-27T03:45:48.121Z8 createdAt: 2026-09-27T03:45:48.121Z
9-updatedAt: 2026-09-29T06:25:13.129Z9+updatedAt: 2026-10-06T03:59:41.259Z
10 labels: []10 labels: []
11 milestone: v0.3.011 milestone: v0.3.0
12+epic: Agent Memory
12 body: ""13 body: ""
13 tasks: []14 tasks: []
14 comments: []15 comments: []
modified .tickets/issues/0050-create-a-web-site.yaml +2 -2
@@ -1,12 +1,12 @@
11 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
22 id: 50
33 title: Create a web site
4-state: open
4+state: closed
55 author:
66 name: k33g
77 email: ph.charriere@gmail.com
88 createdAt: 2026-09-28T18:44:59.825Z
9-updatedAt: 2026-09-29T06:27:12.181Z
9+updatedAt: 2026-10-06T03:58:01.122Z
1010 labels: []
1111 milestone: v0.3.0
1212 body: ""
@@ -1,12 +1,12 @@
1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
2 id: 502 id: 50
3 title: Create a web site3 title: Create a web site
4-state: open4+state: closed
5 author:5 author:
6 name: k33g6 name: k33g
7 email: ph.charriere@gmail.com7 email: ph.charriere@gmail.com
8 createdAt: 2026-09-28T18:44:59.825Z8 createdAt: 2026-09-28T18:44:59.825Z
9-updatedAt: 2026-09-29T06:27:12.181Z9+updatedAt: 2026-10-06T03:58:01.122Z
10 labels: []10 labels: []
11 milestone: v0.3.011 milestone: v0.3.0
12 body: ""12 body: ""
modified .tickets/issues/0052-buitin-acp-tool-to-query-other-agent-like-mini-m.yaml +2 -2
@@ -1,12 +1,12 @@
11 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
22 id: 52
33 title: Buitin ACP tool to query other agent (like mini me, but not only)
4-state: open
4+state: closed
55 author:
66 name: k33g
77 email: ph.charriere@gmail.com
88 createdAt: 2026-09-29T05:18:13.897Z
9-updatedAt: 2026-10-02T21:15:00.000Z
9+updatedAt: 2026-10-06T03:55:00.084Z
1010 labels:
1111 - priority::high
1212 - kind::feature
@@ -1,12 +1,12 @@
1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
2 id: 522 id: 52
3 title: Buitin ACP tool to query other agent (like mini me, but not only)3 title: Buitin ACP tool to query other agent (like mini me, but not only)
4-state: open4+state: closed
5 author:5 author:
6 name: k33g6 name: k33g
7 email: ph.charriere@gmail.com7 email: ph.charriere@gmail.com
8 createdAt: 2026-09-29T05:18:13.897Z8 createdAt: 2026-09-29T05:18:13.897Z
9-updatedAt: 2026-10-02T21:15:00.000Z9+updatedAt: 2026-10-06T03:55:00.084Z
10 labels:10 labels:
11 - priority::high11 - priority::high
12 - kind::feature12 - kind::feature
modified .tickets/issues/0059-how-to-start-in-yolo-mode.yaml +63 -29
@@ -1,12 +1,12 @@
11 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
22 id: 59
33 title: How to start in yolo mode?
4-state: open
4+state: closed
55 author:
66 name: k33g
77 email: ph.charriere@gmail.com
88 createdAt: 2026-10-01T04:39:53.878Z
9-updatedAt: 2026-10-03T10:00:00.000Z
9+updatedAt: 2026-10-06T03:53:27.047Z
1010 labels:
1111 - kind::feature
1212 - priority::low
@@ -50,104 +50,104 @@ body: |
5050 - Make it visible: banner/status line "yolo" so nobody runs it by mistake.
5151 tasks:
5252 - id: 1
53- title: "Decide the open questions below (flag name, ACP scope, MCP scope, banner)"
54- state: open
53+ title: Decide the open questions below (flag name, ACP scope, MCP scope, banner)
54+ state: closed
5555 priority: none
5656 author:
5757 name: k33g
5858 email: ph.charriere@gmail.com
5959 createdAt: 2026-10-01T06:30:00.000Z
60- updatedAt: 2026-10-01T06:30:00.000Z
60+ updatedAt: 2026-10-06T12:00:00.000Z
6161 - id: 2
6262 title: "Failing tests first: with yolo on, window mode (permissions: ask) and ACP Allow() return true without asking; MCP ask-tools run"
63- state: open
63+ state: closed
6464 priority: none
6565 author:
6666 name: k33g
6767 email: ph.charriere@gmail.com
6868 createdAt: 2026-10-01T06:30:00.000Z
69- updatedAt: 2026-10-01T06:30:00.000Z
69+ updatedAt: 2026-10-06T12:00:00.000Z
7070 - id: 3
71- title: "Flag --yolo (main.go, alongside -p/-tui/-acp) + optional env/config override; forces Permissions=auto and mcp.permissions=auto, per-server MCP permissions included"
72- state: open
71+ title: Flag --yolo (main.go, alongside -p/-tui/-acp) + optional env/config override; forces Permissions=auto and mcp.permissions=auto, per-server MCP permissions included
72+ state: closed
7373 priority: none
7474 author:
7575 name: k33g
7676 email: ph.charriere@gmail.com
7777 createdAt: 2026-10-01T06:30:00.000Z
78- updatedAt: 2026-10-01T06:30:00.000Z
78+ updatedAt: 2026-10-06T12:00:00.000Z
7979 - id: 4
8080 title: "ACP sink Allow(): short-circuit when yolo is on (no RequestPermission sent to the editor)"
81- state: open
81+ state: closed
8282 priority: none
8383 author:
8484 name: k33g
8585 email: ph.charriere@gmail.com
8686 createdAt: 2026-10-01T06:30:00.000Z
87- updatedAt: 2026-10-01T06:30:00.000Z
87+ updatedAt: 2026-10-06T12:00:00.000Z
8888 - id: 5
89- title: "Banner/status line shows that yolo is on (terminal, window, ACP agent info)"
90- state: open
89+ title: Banner/status line shows that yolo is on (terminal, window, ACP agent info)
90+ state: closed
9191 priority: none
9292 author:
9393 name: k33g
9494 email: ph.charriere@gmail.com
9595 createdAt: 2026-10-01T06:30:00.000Z
96- updatedAt: 2026-10-01T06:30:00.000Z
96+ updatedAt: 2026-10-06T12:00:00.000Z
9797 - id: 6
9898 title: "Docs EN+FR: reference/cli.md, reference/configuration.md, how-to (window mode, editor); answer the question in the how-to"
99- state: open
99+ state: closed
100100 priority: none
101101 author:
102102 name: k33g
103103 email: ph.charriere@gmail.com
104104 createdAt: 2026-10-01T06:30:00.000Z
105- updatedAt: 2026-10-01T06:30:00.000Z
105+ updatedAt: 2026-10-06T12:00:00.000Z
106106 - id: 7
107107 title: "Per-tool permission: bashTool/editTools accept true | false | ask (config validation, defaults unchanged)"
108- state: open
108+ state: closed
109109 priority: none
110110 author:
111111 name: k33g
112112 email: ph.charriere@gmail.com
113113 createdAt: 2026-10-03T10:00:00.000Z
114- updatedAt: 2026-10-03T10:00:00.000Z
114+ updatedAt: 2026-10-06T12:00:00.000Z
115115 - id: 8
116116 title: "Line mode: a terminal permission prompt (y once / s session / n reject) so `ask` works there too"
117- state: open
117+ state: closed
118118 priority: none
119119 author:
120120 name: k33g
121121 email: ph.charriere@gmail.com
122122 createdAt: 2026-10-03T10:00:00.000Z
123- updatedAt: 2026-10-03T10:00:00.000Z
123+ updatedAt: 2026-10-06T12:00:00.000Z
124124 - id: 9
125- title: "Decide precedence between the per-tool value, the global `permissions` key and --yolo"
126- state: open
125+ title: Decide precedence between the per-tool value, the global `permissions` key and --yolo
126+ state: closed
127127 priority: none
128128 author:
129129 name: k33g
130130 email: ph.charriere@gmail.com
131131 createdAt: 2026-10-03T10:00:00.000Z
132- updatedAt: 2026-10-03T10:00:00.000Z
132+ updatedAt: 2026-10-06T12:00:00.000Z
133133 - id: 10
134- title: "Ship `permissions:` (commented, explained) in agent.*.yaml, demo*/ and config.samples/"
135- state: open
134+ title: Ship `permissions:` (commented, explained) in agent.*.yaml, demo*/ and config.samples/
135+ state: closed
136136 priority: none
137137 author:
138138 name: k33g
139139 email: ph.charriere@gmail.com
140140 createdAt: 2026-10-03T10:00:00.000Z
141- updatedAt: 2026-10-03T10:00:00.000Z
141+ updatedAt: 2026-10-06T12:00:00.000Z
142142 - id: 11
143143 title: "mm --ask: let the caller forbid or confirm the asked agent's tools (ticket 73 open task)"
144- state: open
144+ state: closed
145145 priority: none
146146 author:
147147 name: k33g
148148 email: ph.charriere@gmail.com
149149 createdAt: 2026-10-03T10:00:00.000Z
150- updatedAt: 2026-10-03T10:00:00.000Z
150+ updatedAt: 2026-10-06T12:00:00.000Z
151151 comments:
152152 - id: 1
153153 author:
@@ -195,3 +195,37 @@ comments:
195195 1. Precedence: per-tool value > `permissions` > default — OK?
196196 2. In -p mode, should `ask` reject (safe, proposed) or allow?
197197 3. Should `ask` for editTools cover read_file too (today never gated)?
198+ - id: 2
199+ author:
200+ name: claude
201+ email: noreply@anthropic.com
202+ createdAt: 2026-10-06T12:00:00.000Z
203+ body: |
204+ **Implemented on branch `feature/yolo` (2026-10-06, with ticket 13), uncommitted.**
205+
206+ User's decisions: a config key, not a flag; yolo everywhere ACP
207+ included; `yolo` replaces `permissions` (still read, deprecated);
208+ `yolo: false` forces MCP to ask; four answers (once / session /
209+ reject / never); the agent that holds mmConfigs wins.
210+
211+ - `yolo: true|false` (default true), override `AGENT_YOLO`.
212+ `bashTool`/`editTools` unchanged: they say whether a tool exists.
213+ - `permissions: ask` = `yolo: false` (+ banner warning), `auto` =
214+ warning only. Answers question 9: there is no per-tool value.
215+ - New package `internal/consent`: `Gate` (yolo check), `Answer`,
216+ `Policy` (session memory per tool), `SetPrompter`/`Terminal`/
217+ `Prompt` for the line mode. All six built-in call sites use it.
218+ - Line mode: `🔐` prompt routed by the REPL waiting room
219+ (`internal/agent/prompt.go`), also for /ask. `-p` and `--ask`
220+ reject with a stderr notice (open question 2: reject).
221+ - Window: a/s/r/n. ACP: 4th option `never` (reject_always); with
222+ yolo on no RequestPermission is sent (ticket 57's pending-dialog
223+ case cannot happen in yolo mode).
224+ - MCP: asks when `permissions: ask` OR yolo off; the terminal path
225+ now asks too.
226+ - /ask children run with `AGENT_YOLO=false`; the asker's
227+ `client.RequestPermission` goes through ITS `consent.Gate`.
228+ - Banner: `| yolo: on` / `| yolo: off (tools ask)`.
229+ - Open question 3 (read_file gated?): no, reads never ask.
230+ Docs EN+FR: how-to/confirm-tool-calls.md, explanation/yolo-and-permissions.md,
231+ reference configuration/cli/acp/tools/slash-commands updated.
@@ -1,12 +1,12 @@
1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
2 id: 592 id: 59
3 title: How to start in yolo mode?3 title: How to start in yolo mode?
4-state: open4+state: closed
5 author:5 author:
6 name: k33g6 name: k33g
7 email: ph.charriere@gmail.com7 email: ph.charriere@gmail.com
8 createdAt: 2026-10-01T04:39:53.878Z8 createdAt: 2026-10-01T04:39:53.878Z
9-updatedAt: 2026-10-03T10:00:00.000Z9+updatedAt: 2026-10-06T03:53:27.047Z
10 labels:10 labels:
11 - kind::feature11 - kind::feature
12 - priority::low12 - priority::low
@@ -50,104 +50,104 @@ body: |
50 - Make it visible: banner/status line "yolo" so nobody runs it by mistake.50 - Make it visible: banner/status line "yolo" so nobody runs it by mistake.
51 tasks:51 tasks:
52 - id: 152 - id: 1
53- title: "Decide the open questions below (flag name, ACP scope, MCP scope, banner)"53+ title: Decide the open questions below (flag name, ACP scope, MCP scope, banner)
54- state: open54+ state: closed
55 priority: none55 priority: none
56 author:56 author:
57 name: k33g57 name: k33g
58 email: ph.charriere@gmail.com58 email: ph.charriere@gmail.com
59 createdAt: 2026-10-01T06:30:00.000Z59 createdAt: 2026-10-01T06:30:00.000Z
60- updatedAt: 2026-10-01T06:30:00.000Z60+ updatedAt: 2026-10-06T12:00:00.000Z
61 - id: 261 - id: 2
62 title: "Failing tests first: with yolo on, window mode (permissions: ask) and ACP Allow() return true without asking; MCP ask-tools run"62 title: "Failing tests first: with yolo on, window mode (permissions: ask) and ACP Allow() return true without asking; MCP ask-tools run"
63- state: open63+ state: closed
64 priority: none64 priority: none
65 author:65 author:
66 name: k33g66 name: k33g
67 email: ph.charriere@gmail.com67 email: ph.charriere@gmail.com
68 createdAt: 2026-10-01T06:30:00.000Z68 createdAt: 2026-10-01T06:30:00.000Z
69- updatedAt: 2026-10-01T06:30:00.000Z69+ updatedAt: 2026-10-06T12:00:00.000Z
70 - id: 370 - id: 3
71- title: "Flag --yolo (main.go, alongside -p/-tui/-acp) + optional env/config override; forces Permissions=auto and mcp.permissions=auto, per-server MCP permissions included"71+ title: Flag --yolo (main.go, alongside -p/-tui/-acp) + optional env/config override; forces Permissions=auto and mcp.permissions=auto, per-server MCP permissions included
72- state: open72+ state: closed
73 priority: none73 priority: none
74 author:74 author:
75 name: k33g75 name: k33g
76 email: ph.charriere@gmail.com76 email: ph.charriere@gmail.com
77 createdAt: 2026-10-01T06:30:00.000Z77 createdAt: 2026-10-01T06:30:00.000Z
78- updatedAt: 2026-10-01T06:30:00.000Z78+ updatedAt: 2026-10-06T12:00:00.000Z
79 - id: 479 - id: 4
80 title: "ACP sink Allow(): short-circuit when yolo is on (no RequestPermission sent to the editor)"80 title: "ACP sink Allow(): short-circuit when yolo is on (no RequestPermission sent to the editor)"
81- state: open81+ state: closed
82 priority: none82 priority: none
83 author:83 author:
84 name: k33g84 name: k33g
85 email: ph.charriere@gmail.com85 email: ph.charriere@gmail.com
86 createdAt: 2026-10-01T06:30:00.000Z86 createdAt: 2026-10-01T06:30:00.000Z
87- updatedAt: 2026-10-01T06:30:00.000Z87+ updatedAt: 2026-10-06T12:00:00.000Z
88 - id: 588 - id: 5
89- title: "Banner/status line shows that yolo is on (terminal, window, ACP agent info)"89+ title: Banner/status line shows that yolo is on (terminal, window, ACP agent info)
90- state: open90+ state: closed
91 priority: none91 priority: none
92 author:92 author:
93 name: k33g93 name: k33g
94 email: ph.charriere@gmail.com94 email: ph.charriere@gmail.com
95 createdAt: 2026-10-01T06:30:00.000Z95 createdAt: 2026-10-01T06:30:00.000Z
96- updatedAt: 2026-10-01T06:30:00.000Z96+ updatedAt: 2026-10-06T12:00:00.000Z
97 - id: 697 - id: 6
98 title: "Docs EN+FR: reference/cli.md, reference/configuration.md, how-to (window mode, editor); answer the question in the how-to"98 title: "Docs EN+FR: reference/cli.md, reference/configuration.md, how-to (window mode, editor); answer the question in the how-to"
99- state: open99+ state: closed
100 priority: none100 priority: none
101 author:101 author:
102 name: k33g102 name: k33g
103 email: ph.charriere@gmail.com103 email: ph.charriere@gmail.com
104 createdAt: 2026-10-01T06:30:00.000Z104 createdAt: 2026-10-01T06:30:00.000Z
105- updatedAt: 2026-10-01T06:30:00.000Z105+ updatedAt: 2026-10-06T12:00:00.000Z
106 - id: 7106 - id: 7
107 title: "Per-tool permission: bashTool/editTools accept true | false | ask (config validation, defaults unchanged)"107 title: "Per-tool permission: bashTool/editTools accept true | false | ask (config validation, defaults unchanged)"
108- state: open108+ state: closed
109 priority: none109 priority: none
110 author:110 author:
111 name: k33g111 name: k33g
112 email: ph.charriere@gmail.com112 email: ph.charriere@gmail.com
113 createdAt: 2026-10-03T10:00:00.000Z113 createdAt: 2026-10-03T10:00:00.000Z
114- updatedAt: 2026-10-03T10:00:00.000Z114+ updatedAt: 2026-10-06T12:00:00.000Z
115 - id: 8115 - id: 8
116 title: "Line mode: a terminal permission prompt (y once / s session / n reject) so `ask` works there too"116 title: "Line mode: a terminal permission prompt (y once / s session / n reject) so `ask` works there too"
117- state: open117+ state: closed
118 priority: none118 priority: none
119 author:119 author:
120 name: k33g120 name: k33g
121 email: ph.charriere@gmail.com121 email: ph.charriere@gmail.com
122 createdAt: 2026-10-03T10:00:00.000Z122 createdAt: 2026-10-03T10:00:00.000Z
123- updatedAt: 2026-10-03T10:00:00.000Z123+ updatedAt: 2026-10-06T12:00:00.000Z
124 - id: 9124 - id: 9
125- title: "Decide precedence between the per-tool value, the global `permissions` key and --yolo"125+ title: Decide precedence between the per-tool value, the global `permissions` key and --yolo
126- state: open126+ state: closed
127 priority: none127 priority: none
128 author:128 author:
129 name: k33g129 name: k33g
130 email: ph.charriere@gmail.com130 email: ph.charriere@gmail.com
131 createdAt: 2026-10-03T10:00:00.000Z131 createdAt: 2026-10-03T10:00:00.000Z
132- updatedAt: 2026-10-03T10:00:00.000Z132+ updatedAt: 2026-10-06T12:00:00.000Z
133 - id: 10133 - id: 10
134- title: "Ship `permissions:` (commented, explained) in agent.*.yaml, demo*/ and config.samples/"134+ title: Ship `permissions:` (commented, explained) in agent.*.yaml, demo*/ and config.samples/
135- state: open135+ state: closed
136 priority: none136 priority: none
137 author:137 author:
138 name: k33g138 name: k33g
139 email: ph.charriere@gmail.com139 email: ph.charriere@gmail.com
140 createdAt: 2026-10-03T10:00:00.000Z140 createdAt: 2026-10-03T10:00:00.000Z
141- updatedAt: 2026-10-03T10:00:00.000Z141+ updatedAt: 2026-10-06T12:00:00.000Z
142 - id: 11142 - id: 11
143 title: "mm --ask: let the caller forbid or confirm the asked agent's tools (ticket 73 open task)"143 title: "mm --ask: let the caller forbid or confirm the asked agent's tools (ticket 73 open task)"
144- state: open144+ state: closed
145 priority: none145 priority: none
146 author:146 author:
147 name: k33g147 name: k33g
148 email: ph.charriere@gmail.com148 email: ph.charriere@gmail.com
149 createdAt: 2026-10-03T10:00:00.000Z149 createdAt: 2026-10-03T10:00:00.000Z
150- updatedAt: 2026-10-03T10:00:00.000Z150+ updatedAt: 2026-10-06T12:00:00.000Z
151 comments:151 comments:
152 - id: 1152 - id: 1
153 author:153 author:
@@ -195,3 +195,37 @@ comments:
195 1. Precedence: per-tool value > `permissions` > default — OK?195 1. Precedence: per-tool value > `permissions` > default — OK?
196 2. In -p mode, should `ask` reject (safe, proposed) or allow?196 2. In -p mode, should `ask` reject (safe, proposed) or allow?
197 3. Should `ask` for editTools cover read_file too (today never gated)?197 3. Should `ask` for editTools cover read_file too (today never gated)?
198+ - id: 2
199+ author:
200+ name: claude
201+ email: noreply@anthropic.com
202+ createdAt: 2026-10-06T12:00:00.000Z
203+ body: |
204+ **Implemented on branch `feature/yolo` (2026-10-06, with ticket 13), uncommitted.**
205+
206+ User's decisions: a config key, not a flag; yolo everywhere ACP
207+ included; `yolo` replaces `permissions` (still read, deprecated);
208+ `yolo: false` forces MCP to ask; four answers (once / session /
209+ reject / never); the agent that holds mmConfigs wins.
210+
211+ - `yolo: true|false` (default true), override `AGENT_YOLO`.
212+ `bashTool`/`editTools` unchanged: they say whether a tool exists.
213+ - `permissions: ask` = `yolo: false` (+ banner warning), `auto` =
214+ warning only. Answers question 9: there is no per-tool value.
215+ - New package `internal/consent`: `Gate` (yolo check), `Answer`,
216+ `Policy` (session memory per tool), `SetPrompter`/`Terminal`/
217+ `Prompt` for the line mode. All six built-in call sites use it.
218+ - Line mode: `🔐` prompt routed by the REPL waiting room
219+ (`internal/agent/prompt.go`), also for /ask. `-p` and `--ask`
220+ reject with a stderr notice (open question 2: reject).
221+ - Window: a/s/r/n. ACP: 4th option `never` (reject_always); with
222+ yolo on no RequestPermission is sent (ticket 57's pending-dialog
223+ case cannot happen in yolo mode).
224+ - MCP: asks when `permissions: ask` OR yolo off; the terminal path
225+ now asks too.
226+ - /ask children run with `AGENT_YOLO=false`; the asker's
227+ `client.RequestPermission` goes through ITS `consent.Gate`.
228+ - Banner: `| yolo: on` / `| yolo: off (tools ask)`.
229+ - Open question 3 (read_file gated?): no, reads never ask.
230+ Docs EN+FR: how-to/confirm-tool-calls.md, explanation/yolo-and-permissions.md,
231+ reference configuration/cli/acp/tools/slash-commands updated.
modified .tickets/issues/0060-add-llmman-provider.yaml +8 -8
@@ -1,12 +1,12 @@
11 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
22 id: 60
33 title: Add llmman provider
4-state: open
4+state: closed
55 author:
66 name: k33g
77 email: ph.charriere@gmail.com
88 createdAt: 2026-10-01T04:48:02.829Z
9-updatedAt: 2026-10-01T07:00:00.000Z
9+updatedAt: 2026-10-06T03:56:22.404Z
1010 labels:
1111 - kind::feature
1212 - priority::medium
@@ -19,7 +19,7 @@ body: |
1919 demo-llmman/ using hf.co/unsloth/Qwen3.5-4B-GGUF:Q4_K_M.
2020 tasks:
2121 - id: 1
22- title: "Registry entry llmman (URL, fallback, key, progress, thinking off)"
22+ title: Registry entry llmman (URL, fallback, key, progress, thinking off)
2323 state: closed
2424 priority: none
2525 author:
@@ -46,7 +46,7 @@ tasks:
4646 createdAt: 2026-10-01T07:00:00.000Z
4747 updatedAt: 2026-10-01T07:00:00.000Z
4848 - id: 4
49- title: "demo-llmman/agent.llmman.yaml with hf.co/unsloth/Qwen3.5-4B-GGUF:Q4_K_M"
49+ title: demo-llmman/agent.llmman.yaml with hf.co/unsloth/Qwen3.5-4B-GGUF:Q4_K_M
5050 state: closed
5151 priority: none
5252 author:
@@ -55,7 +55,7 @@ tasks:
5555 createdAt: 2026-10-01T07:00:00.000Z
5656 updatedAt: 2026-10-01T07:00:00.000Z
5757 - id: 5
58- title: "Docs EN+FR (how-to, reference, explanation, indexes)"
58+ title: Docs EN+FR (how-to, reference, explanation, indexes)
5959 state: closed
6060 priority: none
6161 author:
@@ -64,7 +64,7 @@ tasks:
6464 createdAt: 2026-10-01T07:00:00.000Z
6565 updatedAt: 2026-10-01T07:00:00.000Z
6666 - id: 6
67- title: "Live check against llmman serve (tool call, ctx, compaction)"
67+ title: Live check against llmman serve (tool call, ctx, compaction)
6868 state: closed
6969 priority: none
7070 author:
@@ -73,7 +73,7 @@ tasks:
7373 createdAt: 2026-10-01T07:00:00.000Z
7474 updatedAt: 2026-10-01T07:00:00.000Z
7575 - id: 7
76- title: "Quality gate (qlty) where available"
76+ title: Quality gate (qlty) where available
7777 state: open
7878 priority: none
7979 author:
@@ -82,7 +82,7 @@ tasks:
8282 createdAt: 2026-10-01T07:00:00.000Z
8383 updatedAt: 2026-10-01T07:00:00.000Z
8484 - id: 8
85- title: "Review, commit, merge"
85+ title: Review, commit, merge
8686 state: open
8787 priority: none
8888 author:
@@ -1,12 +1,12 @@
1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
2 id: 602 id: 60
3 title: Add llmman provider3 title: Add llmman provider
4-state: open4+state: closed
5 author:5 author:
6 name: k33g6 name: k33g
7 email: ph.charriere@gmail.com7 email: ph.charriere@gmail.com
8 createdAt: 2026-10-01T04:48:02.829Z8 createdAt: 2026-10-01T04:48:02.829Z
9-updatedAt: 2026-10-01T07:00:00.000Z9+updatedAt: 2026-10-06T03:56:22.404Z
10 labels:10 labels:
11 - kind::feature11 - kind::feature
12 - priority::medium12 - priority::medium
@@ -19,7 +19,7 @@ body: |
19 demo-llmman/ using hf.co/unsloth/Qwen3.5-4B-GGUF:Q4_K_M.19 demo-llmman/ using hf.co/unsloth/Qwen3.5-4B-GGUF:Q4_K_M.
20 tasks:20 tasks:
21 - id: 121 - id: 1
22- title: "Registry entry llmman (URL, fallback, key, progress, thinking off)"22+ title: Registry entry llmman (URL, fallback, key, progress, thinking off)
23 state: closed23 state: closed
24 priority: none24 priority: none
25 author:25 author:
@@ -46,7 +46,7 @@ tasks:
46 createdAt: 2026-10-01T07:00:00.000Z46 createdAt: 2026-10-01T07:00:00.000Z
47 updatedAt: 2026-10-01T07:00:00.000Z47 updatedAt: 2026-10-01T07:00:00.000Z
48 - id: 448 - id: 4
49- title: "demo-llmman/agent.llmman.yaml with hf.co/unsloth/Qwen3.5-4B-GGUF:Q4_K_M"49+ title: demo-llmman/agent.llmman.yaml with hf.co/unsloth/Qwen3.5-4B-GGUF:Q4_K_M
50 state: closed50 state: closed
51 priority: none51 priority: none
52 author:52 author:
@@ -55,7 +55,7 @@ tasks:
55 createdAt: 2026-10-01T07:00:00.000Z55 createdAt: 2026-10-01T07:00:00.000Z
56 updatedAt: 2026-10-01T07:00:00.000Z56 updatedAt: 2026-10-01T07:00:00.000Z
57 - id: 557 - id: 5
58- title: "Docs EN+FR (how-to, reference, explanation, indexes)"58+ title: Docs EN+FR (how-to, reference, explanation, indexes)
59 state: closed59 state: closed
60 priority: none60 priority: none
61 author:61 author:
@@ -64,7 +64,7 @@ tasks:
64 createdAt: 2026-10-01T07:00:00.000Z64 createdAt: 2026-10-01T07:00:00.000Z
65 updatedAt: 2026-10-01T07:00:00.000Z65 updatedAt: 2026-10-01T07:00:00.000Z
66 - id: 666 - id: 6
67- title: "Live check against llmman serve (tool call, ctx, compaction)"67+ title: Live check against llmman serve (tool call, ctx, compaction)
68 state: closed68 state: closed
69 priority: none69 priority: none
70 author:70 author:
@@ -73,7 +73,7 @@ tasks:
73 createdAt: 2026-10-01T07:00:00.000Z73 createdAt: 2026-10-01T07:00:00.000Z
74 updatedAt: 2026-10-01T07:00:00.000Z74 updatedAt: 2026-10-01T07:00:00.000Z
75 - id: 775 - id: 7
76- title: "Quality gate (qlty) where available"76+ title: Quality gate (qlty) where available
77 state: open77 state: open
78 priority: none78 priority: none
79 author:79 author:
@@ -82,7 +82,7 @@ tasks:
82 createdAt: 2026-10-01T07:00:00.000Z82 createdAt: 2026-10-01T07:00:00.000Z
83 updatedAt: 2026-10-01T07:00:00.000Z83 updatedAt: 2026-10-01T07:00:00.000Z
84 - id: 884 - id: 8
85- title: "Review, commit, merge"85+ title: Review, commit, merge
86 state: open86 state: open
87 priority: none87 priority: none
88 author:88 author:
modified .tickets/issues/0062-add-ollama-provider.yaml +11 -11
@@ -1,12 +1,12 @@
11 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
22 id: 62
33 title: Add Ollama provider
4-state: open
4+state: closed
55 author:
66 name: k33g
77 email: ph.charriere@gmail.com
88 createdAt: 2026-10-01T06:10:07.295Z
9-updatedAt: 2026-10-01T07:00:00.000Z
9+updatedAt: 2026-10-06T03:56:14.111Z
1010 labels:
1111 - kind::feature
1212 - priority::medium
@@ -19,7 +19,7 @@ body: |
1919 later is one field to change.
2020 tasks:
2121 - id: 1
22- title: "Measure Ollama's /v1 (reasoning field, progress, thinking switch, window, 404)"
22+ title: Measure Ollama's /v1 (reasoning field, progress, thinking switch, window, 404)
2323 state: closed
2424 priority: none
2525 author:
@@ -28,7 +28,7 @@ tasks:
2828 createdAt: 2026-10-01T07:00:00.000Z
2929 updatedAt: 2026-10-01T07:00:00.000Z
3030 - id: 2
31- title: "Per-provider thinkingOff field; reasoning_effort none for Ollama"
31+ title: Per-provider thinkingOff field; reasoning_effort none for Ollama
3232 state: closed
3333 priority: none
3434 author:
@@ -37,7 +37,7 @@ tasks:
3737 createdAt: 2026-10-01T07:00:00.000Z
3838 updatedAt: 2026-10-01T07:00:00.000Z
3939 - id: 3
40- title: "Registry entry ollama + probeOllama (shared /api/ps reading with llmman)"
40+ title: Registry entry ollama + probeOllama (shared /api/ps reading with llmman)
4141 state: closed
4242 priority: none
4343 author:
@@ -55,7 +55,7 @@ tasks:
5555 createdAt: 2026-10-01T07:00:00.000Z
5656 updatedAt: 2026-10-01T07:00:00.000Z
5757 - id: 5
58- title: "demo-ollama/agent.ollama.yaml with qwen3.5:2b"
58+ title: demo-ollama/agent.ollama.yaml with qwen3.5:2b
5959 state: closed
6060 priority: none
6161 author:
@@ -64,7 +64,7 @@ tasks:
6464 createdAt: 2026-10-01T07:00:00.000Z
6565 updatedAt: 2026-10-01T07:00:00.000Z
6666 - id: 6
67- title: "Docs EN+FR (how-to, reference, explanation, indexes)"
67+ title: Docs EN+FR (how-to, reference, explanation, indexes)
6868 state: closed
6969 priority: none
7070 author:
@@ -73,7 +73,7 @@ tasks:
7373 createdAt: 2026-10-01T07:00:00.000Z
7474 updatedAt: 2026-10-01T07:00:00.000Z
7575 - id: 7
76- title: "Live check against Ollama (tool call, ctx, compaction)"
76+ title: Live check against Ollama (tool call, ctx, compaction)
7777 state: closed
7878 priority: none
7979 author:
@@ -82,7 +82,7 @@ tasks:
8282 createdAt: 2026-10-01T07:00:00.000Z
8383 updatedAt: 2026-10-01T07:00:00.000Z
8484 - id: 8
85- title: "Check reasoning_effort none on a non-thinking model"
85+ title: Check reasoning_effort none on a non-thinking model
8686 state: open
8787 priority: none
8888 author:
@@ -91,7 +91,7 @@ tasks:
9191 createdAt: 2026-10-01T07:00:00.000Z
9292 updatedAt: 2026-10-01T07:00:00.000Z
9393 - id: 9
94- title: "Quality gate (qlty) where available"
94+ title: Quality gate (qlty) where available
9595 state: open
9696 priority: none
9797 author:
@@ -100,7 +100,7 @@ tasks:
100100 createdAt: 2026-10-01T07:00:00.000Z
101101 updatedAt: 2026-10-01T07:00:00.000Z
102102 - id: 10
103- title: "Review, commit, merge"
103+ title: Review, commit, merge
104104 state: open
105105 priority: none
106106 author:
@@ -1,12 +1,12 @@
1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
2 id: 622 id: 62
3 title: Add Ollama provider3 title: Add Ollama provider
4-state: open4+state: closed
5 author:5 author:
6 name: k33g6 name: k33g
7 email: ph.charriere@gmail.com7 email: ph.charriere@gmail.com
8 createdAt: 2026-10-01T06:10:07.295Z8 createdAt: 2026-10-01T06:10:07.295Z
9-updatedAt: 2026-10-01T07:00:00.000Z9+updatedAt: 2026-10-06T03:56:14.111Z
10 labels:10 labels:
11 - kind::feature11 - kind::feature
12 - priority::medium12 - priority::medium
@@ -19,7 +19,7 @@ body: |
19 later is one field to change.19 later is one field to change.
20 tasks:20 tasks:
21 - id: 121 - id: 1
22- title: "Measure Ollama's /v1 (reasoning field, progress, thinking switch, window, 404)"22+ title: Measure Ollama's /v1 (reasoning field, progress, thinking switch, window, 404)
23 state: closed23 state: closed
24 priority: none24 priority: none
25 author:25 author:
@@ -28,7 +28,7 @@ tasks:
28 createdAt: 2026-10-01T07:00:00.000Z28 createdAt: 2026-10-01T07:00:00.000Z
29 updatedAt: 2026-10-01T07:00:00.000Z29 updatedAt: 2026-10-01T07:00:00.000Z
30 - id: 230 - id: 2
31- title: "Per-provider thinkingOff field; reasoning_effort none for Ollama"31+ title: Per-provider thinkingOff field; reasoning_effort none for Ollama
32 state: closed32 state: closed
33 priority: none33 priority: none
34 author:34 author:
@@ -37,7 +37,7 @@ tasks:
37 createdAt: 2026-10-01T07:00:00.000Z37 createdAt: 2026-10-01T07:00:00.000Z
38 updatedAt: 2026-10-01T07:00:00.000Z38 updatedAt: 2026-10-01T07:00:00.000Z
39 - id: 339 - id: 3
40- title: "Registry entry ollama + probeOllama (shared /api/ps reading with llmman)"40+ title: Registry entry ollama + probeOllama (shared /api/ps reading with llmman)
41 state: closed41 state: closed
42 priority: none42 priority: none
43 author:43 author:
@@ -55,7 +55,7 @@ tasks:
55 createdAt: 2026-10-01T07:00:00.000Z55 createdAt: 2026-10-01T07:00:00.000Z
56 updatedAt: 2026-10-01T07:00:00.000Z56 updatedAt: 2026-10-01T07:00:00.000Z
57 - id: 557 - id: 5
58- title: "demo-ollama/agent.ollama.yaml with qwen3.5:2b"58+ title: demo-ollama/agent.ollama.yaml with qwen3.5:2b
59 state: closed59 state: closed
60 priority: none60 priority: none
61 author:61 author:
@@ -64,7 +64,7 @@ tasks:
64 createdAt: 2026-10-01T07:00:00.000Z64 createdAt: 2026-10-01T07:00:00.000Z
65 updatedAt: 2026-10-01T07:00:00.000Z65 updatedAt: 2026-10-01T07:00:00.000Z
66 - id: 666 - id: 6
67- title: "Docs EN+FR (how-to, reference, explanation, indexes)"67+ title: Docs EN+FR (how-to, reference, explanation, indexes)
68 state: closed68 state: closed
69 priority: none69 priority: none
70 author:70 author:
@@ -73,7 +73,7 @@ tasks:
73 createdAt: 2026-10-01T07:00:00.000Z73 createdAt: 2026-10-01T07:00:00.000Z
74 updatedAt: 2026-10-01T07:00:00.000Z74 updatedAt: 2026-10-01T07:00:00.000Z
75 - id: 775 - id: 7
76- title: "Live check against Ollama (tool call, ctx, compaction)"76+ title: Live check against Ollama (tool call, ctx, compaction)
77 state: closed77 state: closed
78 priority: none78 priority: none
79 author:79 author:
@@ -82,7 +82,7 @@ tasks:
82 createdAt: 2026-10-01T07:00:00.000Z82 createdAt: 2026-10-01T07:00:00.000Z
83 updatedAt: 2026-10-01T07:00:00.000Z83 updatedAt: 2026-10-01T07:00:00.000Z
84 - id: 884 - id: 8
85- title: "Check reasoning_effort none on a non-thinking model"85+ title: Check reasoning_effort none on a non-thinking model
86 state: open86 state: open
87 priority: none87 priority: none
88 author:88 author:
@@ -91,7 +91,7 @@ tasks:
91 createdAt: 2026-10-01T07:00:00.000Z91 createdAt: 2026-10-01T07:00:00.000Z
92 updatedAt: 2026-10-01T07:00:00.000Z92 updatedAt: 2026-10-01T07:00:00.000Z
93 - id: 993 - id: 9
94- title: "Quality gate (qlty) where available"94+ title: Quality gate (qlty) where available
95 state: open95 state: open
96 priority: none96 priority: none
97 author:97 author:
@@ -100,7 +100,7 @@ tasks:
100 createdAt: 2026-10-01T07:00:00.000Z100 createdAt: 2026-10-01T07:00:00.000Z
101 updatedAt: 2026-10-01T07:00:00.000Z101 updatedAt: 2026-10-01T07:00:00.000Z
102 - id: 10102 - id: 10
103- title: "Review, commit, merge"103+ title: Review, commit, merge
104 state: open104 state: open
105 priority: none105 priority: none
106 author:106 author:
modified .tickets/issues/0066-talk-to-another-mm.yaml +8 -8
@@ -1,12 +1,12 @@
11 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
22 id: 66
33 title: talk to another mm
4-state: open
4+state: closed
55 author:
66 name: k33g
77 email: ph.charriere@gmail.com
88 createdAt: 2026-10-02T18:48:36.275Z
9-updatedAt: 2026-10-02T21:15:00.000Z
9+updatedAt: 2026-10-06T03:55:28.931Z
1010 labels:
1111 - kind::feature
1212 related:
@@ -226,7 +226,7 @@ body: |
226226 c. Editor: one /ask with a hint, or one command per agent?
227227 tasks:
228228 - id: 1
229- title: "Answer the open questions (key name, immediate answer, multi-turn, fronts, context passed)"
229+ title: Answer the open questions (key name, immediate answer, multi-turn, fronts, context passed)
230230 state: closed
231231 priority: none
232232 author:
@@ -235,7 +235,7 @@ tasks:
235235 createdAt: 2026-10-02T19:30:00.000Z
236236 updatedAt: 2026-10-02T20:00:00.000Z
237237 - id: 2
238- title: "Confirm the remaining points a/b/c (context format, fresh session per /ask, editor hint vs one command per agent)"
238+ title: Confirm the remaining points a/b/c (context format, fresh session per /ask, editor hint vs one command per agent)
239239 state: closed
240240 priority: none
241241 author:
@@ -289,7 +289,7 @@ tasks:
289289 createdAt: 2026-10-02T19:30:00.000Z
290290 updatedAt: 2026-10-02T20:00:00.000Z
291291 - id: 8
292- title: "Insert the exchange in the current history (format, cap, valid alternation), persisted by the store and replayed on /resume"
292+ title: Insert the exchange in the current history (format, cap, valid alternation), persisted by the store and replayed on /resume
293293 state: closed
294294 priority: none
295295 author:
@@ -298,7 +298,7 @@ tasks:
298298 createdAt: 2026-10-02T19:30:00.000Z
299299 updatedAt: 2026-10-02T20:00:00.000Z
300300 - id: 9
301- title: "/abort, Ctrl+C and editor cancel -> session/cancel on the child; child crash -> one-line error, history untouched"
301+ title: /abort, Ctrl+C and editor cancel -> session/cancel on the child; child crash -> one-line error, history untouched
302302 state: closed
303303 priority: none
304304 author:
@@ -334,7 +334,7 @@ tasks:
334334 createdAt: 2026-10-02T20:30:00.000Z
335335 updatedAt: 2026-10-02T20:30:00.000Z
336336 - id: 13
337- title: "Live check of the window mode (/ask-jan in mm -tui) — unit-tested only"
337+ title: Live check of the window mode (/ask-jan in mm -tui) — unit-tested only
338338 state: open
339339 priority: none
340340 author:
@@ -343,7 +343,7 @@ tasks:
343343 createdAt: 2026-10-02T20:30:00.000Z
344344 updatedAt: 2026-10-02T20:30:00.000Z
345345 - id: 14
346- title: "Review, commit, merge"
346+ title: Review, commit, merge
347347 state: open
348348 priority: none
349349 author:
@@ -1,12 +1,12 @@
1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.1 # Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
2 id: 662 id: 66
3 title: talk to another mm3 title: talk to another mm
4-state: open4+state: closed
5 author:5 author:
6 name: k33g6 name: k33g
7 email: ph.charriere@gmail.com7 email: ph.charriere@gmail.com
8 createdAt: 2026-10-02T18:48:36.275Z8 createdAt: 2026-10-02T18:48:36.275Z
9-updatedAt: 2026-10-02T21:15:00.000Z9+updatedAt: 2026-10-06T03:55:28.931Z
10 labels:10 labels:
11 - kind::feature11 - kind::feature
12 related:12 related:
@@ -226,7 +226,7 @@ body: |
226 c. Editor: one /ask with a hint, or one command per agent?226 c. Editor: one /ask with a hint, or one command per agent?
227 tasks:227 tasks:
228 - id: 1228 - id: 1
229- title: "Answer the open questions (key name, immediate answer, multi-turn, fronts, context passed)"229+ title: Answer the open questions (key name, immediate answer, multi-turn, fronts, context passed)
230 state: closed230 state: closed
231 priority: none231 priority: none
232 author:232 author:
@@ -235,7 +235,7 @@ tasks:
235 createdAt: 2026-10-02T19:30:00.000Z235 createdAt: 2026-10-02T19:30:00.000Z
236 updatedAt: 2026-10-02T20:00:00.000Z236 updatedAt: 2026-10-02T20:00:00.000Z
237 - id: 2237 - id: 2
238- title: "Confirm the remaining points a/b/c (context format, fresh session per /ask, editor hint vs one command per agent)"238+ title: Confirm the remaining points a/b/c (context format, fresh session per /ask, editor hint vs one command per agent)
239 state: closed239 state: closed
240 priority: none240 priority: none
241 author:241 author:
@@ -289,7 +289,7 @@ tasks:
289 createdAt: 2026-10-02T19:30:00.000Z289 createdAt: 2026-10-02T19:30:00.000Z
290 updatedAt: 2026-10-02T20:00:00.000Z290 updatedAt: 2026-10-02T20:00:00.000Z
291 - id: 8291 - id: 8
292- title: "Insert the exchange in the current history (format, cap, valid alternation), persisted by the store and replayed on /resume"292+ title: Insert the exchange in the current history (format, cap, valid alternation), persisted by the store and replayed on /resume
293 state: closed293 state: closed
294 priority: none294 priority: none
295 author:295 author:
@@ -298,7 +298,7 @@ tasks:
298 createdAt: 2026-10-02T19:30:00.000Z298 createdAt: 2026-10-02T19:30:00.000Z
299 updatedAt: 2026-10-02T20:00:00.000Z299 updatedAt: 2026-10-02T20:00:00.000Z
300 - id: 9300 - id: 9
301- title: "/abort, Ctrl+C and editor cancel -> session/cancel on the child; child crash -> one-line error, history untouched"301+ title: /abort, Ctrl+C and editor cancel -> session/cancel on the child; child crash -> one-line error, history untouched
302 state: closed302 state: closed
303 priority: none303 priority: none
304 author:304 author:
@@ -334,7 +334,7 @@ tasks:
334 createdAt: 2026-10-02T20:30:00.000Z334 createdAt: 2026-10-02T20:30:00.000Z
335 updatedAt: 2026-10-02T20:30:00.000Z335 updatedAt: 2026-10-02T20:30:00.000Z
336 - id: 13336 - id: 13
337- title: "Live check of the window mode (/ask-jan in mm -tui) — unit-tested only"337+ title: Live check of the window mode (/ask-jan in mm -tui) — unit-tested only
338 state: open338 state: open
339 priority: none339 priority: none
340 author:340 author:
@@ -343,7 +343,7 @@ tasks:
343 createdAt: 2026-10-02T20:30:00.000Z343 createdAt: 2026-10-02T20:30:00.000Z
344 updatedAt: 2026-10-02T20:30:00.000Z344 updatedAt: 2026-10-02T20:30:00.000Z
345 - id: 14345 - id: 14
346- title: "Review, commit, merge"346+ title: Review, commit, merge
347 state: open347 state: open
348 priority: none348 priority: none
349 author:349 author:
modified .tickets/issues/0067-enable-disable-reasoning-mode.yaml +187 -5
@@ -6,8 +6,190 @@ author:
66 name: k33g
77 email: ph.charriere@gmail.com
88 createdAt: 2026-10-02T06:09:42.599Z
9-updatedAt: 2026-10-02T06:09:42.599Z
10-labels: []
11-body: ""
12-tasks: []
13-comments: []
9+updatedAt: 2026-10-06T14:00:00.000Z
10+labels:
11+ - kind::feature
12+related:
13+ - 62
14+body: |
15+ Question (user, 2026-10-06): can mm turn a model's reasoning mode on or
16+ off?
17+
18+ Answer as of 2026-10-06 (branch feature/yolo, code verified):
19+
20+ 1. HIDE it: yes. `showThinking: false` hides the reasoning in the three
21+ front ends. The model still reasons (same time, same tokens): only
22+ the display changes.
23+
24+ 2. TURN IT OFF for the /compact summary only: done automatically, per
25+ provider (`thinkingOff`, internal/engine/openai_compat.go):
26+ dmr / llamacpp / llmman -> `chat_template_kwargs: {"enable_thinking": false}`,
27+ ollama -> `reasoning_effort: "none"` (ticket 62: Ollama ignores
28+ chat_template_kwargs). Not configurable, never applied to normal turns.
29+
30+ 3. TURN IT OFF for every turn through `sampling`: only half works.
31+ - `sampling.reasoning_effort: none` reaches the server (Ollama should
32+ obey it; not verified live on a normal turn).
33+ - `sampling.chat_template_kwargs: {enable_thinking: false}` is
34+ SILENTLY DROPPED before the request leaves mm, so it is impossible
35+ from the config for dmr / llamacpp / llmman, the servers that need
36+ exactly that field.
37+
38+ ## Why `chat_template_kwargs` is dropped
39+
40+ `chat_template_kwargs` is not an OpenAI API field: it is a llama.cpp
41+ extension (llama-server passes it to the model's Jinja chat template,
42+ where `enable_thinking` switches Qwen3-style thinking off). DMR and
43+ llmman run llama-server underneath, so they expect it too.
44+
45+ The `sampling` map does not go to the server as is. Its road:
46+
47+ 1. internal/engine/engine.go:353 hands it to Genkit:
48+ `ai.WithConfig(config.Cfg.Sampling)`.
49+ 2. Genkit's compat_oai plugin (plugins/compat_oai/generate.go:142,
50+ `WithConfig`) converts the map into the TYPED request struct of
51+ openai-go: `base.MapToStruct[openai.ChatCompletionNewParams](cfg)`.
52+ 3. `MapToStruct` (genkit internal/base/json.go:183) is a JSON round trip:
53+ `json.Marshal(map)` then `json.Unmarshal` into the struct.
54+ 4. `ChatCompletionNewParams` has a field for every OpenAI parameter —
55+ `temperature`, `top_p`, `max_tokens`, `parallel_tool_calls`,
56+ `reasoning_effort` … — and NONE for `chat_template_kwargs`.
57+ `json.Unmarshal` ignores keys that match no field: the key is gone.
58+ 5. The HTTP body is then serialised FROM THE STRUCT, so the server never
59+ sees it.
60+
61+ Measured with a round trip in the module (openai-go v1.8.2):
62+ input `{"temperature":0, "reasoning_effort":"none",
63+ "chat_template_kwargs":{"enable_thinking":false}}`
64+ output `{"temperature":0,"reasoning_effort":"none"}`.
65+
66+ Same reason the summary path does NOT use `sampling`: mm adds the field
67+ in its own openai-go middleware (`streamOptions` -> `withoutReasoning`,
68+ internal/engine/usage.go:83 and :137), which edits the raw HTTP body
69+ AFTER serialisation, where nothing can drop it any more.
70+
71+ Workarounds today (no code change): server side —
72+ `llama-server --reasoning-budget 0` (or `--chat-template-kwargs
73+ '{"enable_thinking":false}'` on recent builds — neither flag verified
74+ in this sandbox), or a model without a thinking mode.
75+
76+ ## Proposal
77+
78+ A config key `thinking: true|false` (default true, override
79+ `AGENT_THINKING`), applied by the SAME middleware to every request when
80+ false: the provider's own way (`thinkingOff`), so it works on
81+ dmr/llamacpp/llmman (chat_template_kwargs) and ollama (reasoning_effort).
82+ A request that already carries the field keeps it (rule already in
83+ `withoutReasoning`). Optional: `/think on|off` to switch during a session.
84+tasks:
85+ - id: 1
86+ title: "Decide: key name (`thinking`?), env override, and whether `/think on|off` is wanted"
87+ state: open
88+ priority: none
89+ author:
90+ name: k33g
91+ email: ph.charriere@gmail.com
92+ createdAt: 2026-10-06T13:00:00.000Z
93+ updatedAt: 2026-10-06T13:00:00.000Z
94+ - id: 2
95+ title: "Failing tests first: with thinking off, every request body carries the provider's field (template / effort); unknown provider = unchanged"
96+ state: open
97+ priority: none
98+ author:
99+ name: k33g
100+ email: ph.charriere@gmail.com
101+ createdAt: 2026-10-06T13:00:00.000Z
102+ updatedAt: 2026-10-06T13:00:00.000Z
103+ - id: 3
104+ title: "Config key + AGENT_THINKING; engine: mark every request (not only summaries) with noThinkKey when off"
105+ state: open
106+ priority: none
107+ author:
108+ name: k33g
109+ email: ph.charriere@gmail.com
110+ createdAt: 2026-10-06T13:00:00.000Z
111+ updatedAt: 2026-10-06T13:00:00.000Z
112+ - id: 4
113+ title: "Banner shows `thinking: off`; `showThinking` stays a display setting"
114+ state: open
115+ priority: none
116+ author:
117+ name: k33g
118+ email: ph.charriere@gmail.com
119+ createdAt: 2026-10-06T13:00:00.000Z
120+ updatedAt: 2026-10-06T13:00:00.000Z
121+ - id: 5
122+ title: "Live check per provider: DMR + llmman (chat_template_kwargs), Ollama (reasoning_effort none on a normal turn)"
123+ state: open
124+ priority: none
125+ author:
126+ name: k33g
127+ email: ph.charriere@gmail.com
128+ createdAt: 2026-10-06T13:00:00.000Z
129+ updatedAt: 2026-10-06T13:00:00.000Z
130+ - id: 6
131+ title: "Docs EN+FR: reference configuration, how-to, explanation/thinking-models (why sampling cannot do it)"
132+ state: open
133+ priority: none
134+ author:
135+ name: k33g
136+ email: ph.charriere@gmail.com
137+ createdAt: 2026-10-06T13:00:00.000Z
138+ updatedAt: 2026-10-06T13:00:00.000Z
139+comments:
140+ - id: 1
141+ author:
142+ name: claude
143+ email: noreply@anthropic.com
144+ createdAt: 2026-10-06T14:00:00.000Z
145+ body: |
146+ **Correction: a newer Genkit fixes the "sampling" road (verified in a
147+ scratch copy, the repo is unchanged — still genkit v1.10.0).**
148+
149+ Latest Genkit Go is v1.13.1 (2026-09-03). Since **v1.12.0**,
150+ compat_oai's `WithConfig` keeps the keys `ChatCompletionNewParams`
151+ does not know and sends them as JSON extras (`SetExtraFields`,
152+ plugins/compat_oai/generate.go ~l.231: "the OpenAI SDK silently
153+ ignores provider-specific fields while unmarshaling, so retain them
154+ as JSON extras"). v1.11.0 does not have it.
155+
156+ BUT bumping alone is not enough — found by testing:
157+ 1. In v1.13.1 `OpenAICompatible.DefineModel` is DEPRECATED and
158+ returns an UNREGISTERED model. mm discards its result
159+ (internal/engine/openai_compat.go:189), so the model actually used
160+ is the one the plugin resolves on the fly (`ResolveAction` ->
161+ `newSDKModel`), whose config schema is the strict reflected SDK
162+ schema: `chat_template_kwargs` is then REJECTED at the action
163+ boundary — "config: Additional property chat_template_kwargs is
164+ not allowed" (and every unknown sampling key would fail the
165+ request, where v1.10.0 dropped it silently).
166+ `ModelOptions.ConfigSchema` does not help: it is set on the model
167+ nobody registered.
168+ 2. Fix: register it — `genkit.RegisterAction(g,
169+ client.DefineModel(...).(*ai.ModelAction))`. DefineModel's model
170+ takes its config untyped, no validation, unknown keys ride to the
171+ wire. (The non-deprecated `NewModel` validates against the strict
172+ schema; it would need an explicit permissive `ConfigSchema`.)
173+
174+ Measured (recording server): v1.13.1 + RegisterAction sends
175+ `"chat_template_kwargs": {"enable_thinking": false}` with mm's own
176+ middleware fields (return_progress, stream_options) intact. Live on
177+ DMR qwen3.5-4b, `-p "What is 17*23?"` with
178+ `sampling.chat_template_kwargs.enable_thinking: false`: current mm =
179+ 4 s, a 💭 reasoning block, then 391; patched mm = under 1 s, no
180+ reasoning, 391.
181+
182+ Upgrade side effects seen: `go get` + `go mod tidy` also bump
183+ dotprompt, invopop/jsonschema v0.14.0, goccy/go-yaml, jsonparser and
184+ add go.yaml.in/yaml/v4; `go test ./...` green (scratch copy; the
185+ skills test failed only because the copy had no shipped YAML);
186+ Genkit now logs `INFO Genkit initialized …` on stderr at start-up
187+ (to silence). Not re-checked after the bump: reasoning.go's tap and
188+ toolargs.go's workarounds (v1.13.1 reads reasoning_content itself
189+ now — generate.go ~l.548), the ollama/llmman paths.
190+
191+ Consequence for the proposal: with the bump, `sampling` alone can
192+ carry `chat_template_kwargs` (llama.cpp family) or
193+ `reasoning_effort` (Ollama) — per-server spelling left to the user.
194+ A `thinking: false` key is still worth it for one spelling across
195+ providers (it would reuse `withoutReasoning`) and for `/think`.
@@ -6,8 +6,190 @@ author:
6 name: k33g6 name: k33g
7 email: ph.charriere@gmail.com7 email: ph.charriere@gmail.com
8 createdAt: 2026-10-02T06:09:42.599Z8 createdAt: 2026-10-02T06:09:42.599Z
9-updatedAt: 2026-10-02T06:09:42.599Z9+updatedAt: 2026-10-06T14:00:00.000Z
10-labels: []10+labels:
11-body: ""11+ - kind::feature
12-tasks: []12+related:
13-comments: []13+ - 62
14+body: |
15+ Question (user, 2026-10-06): can mm turn a model's reasoning mode on or
16+ off?
17+
18+ Answer as of 2026-10-06 (branch feature/yolo, code verified):
19+
20+ 1. HIDE it: yes. `showThinking: false` hides the reasoning in the three
21+ front ends. The model still reasons (same time, same tokens): only
22+ the display changes.
23+
24+ 2. TURN IT OFF for the /compact summary only: done automatically, per
25+ provider (`thinkingOff`, internal/engine/openai_compat.go):
26+ dmr / llamacpp / llmman -> `chat_template_kwargs: {"enable_thinking": false}`,
27+ ollama -> `reasoning_effort: "none"` (ticket 62: Ollama ignores
28+ chat_template_kwargs). Not configurable, never applied to normal turns.
29+
30+ 3. TURN IT OFF for every turn through `sampling`: only half works.
31+ - `sampling.reasoning_effort: none` reaches the server (Ollama should
32+ obey it; not verified live on a normal turn).
33+ - `sampling.chat_template_kwargs: {enable_thinking: false}` is
34+ SILENTLY DROPPED before the request leaves mm, so it is impossible
35+ from the config for dmr / llamacpp / llmman, the servers that need
36+ exactly that field.
37+
38+ ## Why `chat_template_kwargs` is dropped
39+
40+ `chat_template_kwargs` is not an OpenAI API field: it is a llama.cpp
41+ extension (llama-server passes it to the model's Jinja chat template,
42+ where `enable_thinking` switches Qwen3-style thinking off). DMR and
43+ llmman run llama-server underneath, so they expect it too.
44+
45+ The `sampling` map does not go to the server as is. Its road:
46+
47+ 1. internal/engine/engine.go:353 hands it to Genkit:
48+ `ai.WithConfig(config.Cfg.Sampling)`.
49+ 2. Genkit's compat_oai plugin (plugins/compat_oai/generate.go:142,
50+ `WithConfig`) converts the map into the TYPED request struct of
51+ openai-go: `base.MapToStruct[openai.ChatCompletionNewParams](cfg)`.
52+ 3. `MapToStruct` (genkit internal/base/json.go:183) is a JSON round trip:
53+ `json.Marshal(map)` then `json.Unmarshal` into the struct.
54+ 4. `ChatCompletionNewParams` has a field for every OpenAI parameter —
55+ `temperature`, `top_p`, `max_tokens`, `parallel_tool_calls`,
56+ `reasoning_effort` … — and NONE for `chat_template_kwargs`.
57+ `json.Unmarshal` ignores keys that match no field: the key is gone.
58+ 5. The HTTP body is then serialised FROM THE STRUCT, so the server never
59+ sees it.
60+
61+ Measured with a round trip in the module (openai-go v1.8.2):
62+ input `{"temperature":0, "reasoning_effort":"none",
63+ "chat_template_kwargs":{"enable_thinking":false}}`
64+ output `{"temperature":0,"reasoning_effort":"none"}`.
65+
66+ Same reason the summary path does NOT use `sampling`: mm adds the field
67+ in its own openai-go middleware (`streamOptions` -> `withoutReasoning`,
68+ internal/engine/usage.go:83 and :137), which edits the raw HTTP body
69+ AFTER serialisation, where nothing can drop it any more.
70+
71+ Workarounds today (no code change): server side —
72+ `llama-server --reasoning-budget 0` (or `--chat-template-kwargs
73+ '{"enable_thinking":false}'` on recent builds — neither flag verified
74+ in this sandbox), or a model without a thinking mode.
75+
76+ ## Proposal
77+
78+ A config key `thinking: true|false` (default true, override
79+ `AGENT_THINKING`), applied by the SAME middleware to every request when
80+ false: the provider's own way (`thinkingOff`), so it works on
81+ dmr/llamacpp/llmman (chat_template_kwargs) and ollama (reasoning_effort).
82+ A request that already carries the field keeps it (rule already in
83+ `withoutReasoning`). Optional: `/think on|off` to switch during a session.
84+tasks:
85+ - id: 1
86+ title: "Decide: key name (`thinking`?), env override, and whether `/think on|off` is wanted"
87+ state: open
88+ priority: none
89+ author:
90+ name: k33g
91+ email: ph.charriere@gmail.com
92+ createdAt: 2026-10-06T13:00:00.000Z
93+ updatedAt: 2026-10-06T13:00:00.000Z
94+ - id: 2
95+ title: "Failing tests first: with thinking off, every request body carries the provider's field (template / effort); unknown provider = unchanged"
96+ state: open
97+ priority: none
98+ author:
99+ name: k33g
100+ email: ph.charriere@gmail.com
101+ createdAt: 2026-10-06T13:00:00.000Z
102+ updatedAt: 2026-10-06T13:00:00.000Z
103+ - id: 3
104+ title: "Config key + AGENT_THINKING; engine: mark every request (not only summaries) with noThinkKey when off"
105+ state: open
106+ priority: none
107+ author:
108+ name: k33g
109+ email: ph.charriere@gmail.com
110+ createdAt: 2026-10-06T13:00:00.000Z
111+ updatedAt: 2026-10-06T13:00:00.000Z
112+ - id: 4
113+ title: "Banner shows `thinking: off`; `showThinking` stays a display setting"
114+ state: open
115+ priority: none
116+ author:
117+ name: k33g
118+ email: ph.charriere@gmail.com
119+ createdAt: 2026-10-06T13:00:00.000Z
120+ updatedAt: 2026-10-06T13:00:00.000Z
121+ - id: 5
122+ title: "Live check per provider: DMR + llmman (chat_template_kwargs), Ollama (reasoning_effort none on a normal turn)"
123+ state: open
124+ priority: none
125+ author:
126+ name: k33g
127+ email: ph.charriere@gmail.com
128+ createdAt: 2026-10-06T13:00:00.000Z
129+ updatedAt: 2026-10-06T13:00:00.000Z
130+ - id: 6
131+ title: "Docs EN+FR: reference configuration, how-to, explanation/thinking-models (why sampling cannot do it)"
132+ state: open
133+ priority: none
134+ author:
135+ name: k33g
136+ email: ph.charriere@gmail.com
137+ createdAt: 2026-10-06T13:00:00.000Z
138+ updatedAt: 2026-10-06T13:00:00.000Z
139+comments:
140+ - id: 1
141+ author:
142+ name: claude
143+ email: noreply@anthropic.com
144+ createdAt: 2026-10-06T14:00:00.000Z
145+ body: |
146+ **Correction: a newer Genkit fixes the "sampling" road (verified in a
147+ scratch copy, the repo is unchanged — still genkit v1.10.0).**
148+
149+ Latest Genkit Go is v1.13.1 (2026-09-03). Since **v1.12.0**,
150+ compat_oai's `WithConfig` keeps the keys `ChatCompletionNewParams`
151+ does not know and sends them as JSON extras (`SetExtraFields`,
152+ plugins/compat_oai/generate.go ~l.231: "the OpenAI SDK silently
153+ ignores provider-specific fields while unmarshaling, so retain them
154+ as JSON extras"). v1.11.0 does not have it.
155+
156+ BUT bumping alone is not enough — found by testing:
157+ 1. In v1.13.1 `OpenAICompatible.DefineModel` is DEPRECATED and
158+ returns an UNREGISTERED model. mm discards its result
159+ (internal/engine/openai_compat.go:189), so the model actually used
160+ is the one the plugin resolves on the fly (`ResolveAction` ->
161+ `newSDKModel`), whose config schema is the strict reflected SDK
162+ schema: `chat_template_kwargs` is then REJECTED at the action
163+ boundary — "config: Additional property chat_template_kwargs is
164+ not allowed" (and every unknown sampling key would fail the
165+ request, where v1.10.0 dropped it silently).
166+ `ModelOptions.ConfigSchema` does not help: it is set on the model
167+ nobody registered.
168+ 2. Fix: register it — `genkit.RegisterAction(g,
169+ client.DefineModel(...).(*ai.ModelAction))`. DefineModel's model
170+ takes its config untyped, no validation, unknown keys ride to the
171+ wire. (The non-deprecated `NewModel` validates against the strict
172+ schema; it would need an explicit permissive `ConfigSchema`.)
173+
174+ Measured (recording server): v1.13.1 + RegisterAction sends
175+ `"chat_template_kwargs": {"enable_thinking": false}` with mm's own
176+ middleware fields (return_progress, stream_options) intact. Live on
177+ DMR qwen3.5-4b, `-p "What is 17*23?"` with
178+ `sampling.chat_template_kwargs.enable_thinking: false`: current mm =
179+ 4 s, a 💭 reasoning block, then 391; patched mm = under 1 s, no
180+ reasoning, 391.
181+
182+ Upgrade side effects seen: `go get` + `go mod tidy` also bump
183+ dotprompt, invopop/jsonschema v0.14.0, goccy/go-yaml, jsonparser and
184+ add go.yaml.in/yaml/v4; `go test ./...` green (scratch copy; the
185+ skills test failed only because the copy had no shipped YAML);
186+ Genkit now logs `INFO Genkit initialized …` on stderr at start-up
187+ (to silence). Not re-checked after the bump: reasoning.go's tap and
188+ toolargs.go's workarounds (v1.13.1 reads reasoning_content itself
189+ now — generate.go ~l.548), the ollama/llmman paths.
190+
191+ Consequence for the proposal: with the bump, `sampling` alone can
192+ carry `chat_template_kwargs` (llama.cpp family) or
193+ `reasoning_effort` (Ollama) — per-server spelling left to the user.
194+ A `thinking: false` key is still worth it for one spelling across
195+ providers (it would reuse `withoutReasoning`) and for `/think`.
renamed .tickets/issues/0073-mm-ask-ask-with-context-ask-another-agent-from-a.yaml +9 -9
Large diff collapsed to keep the page fast — load it when you need it.
modified .tickets/issues/0074-memory-memgpt-hierarchic-memory.yaml +2 -1
@@ -6,9 +6,10 @@ author:
66 name: k33g
77 email: ph.charriere@gmail.com
88 createdAt: 2026-10-04T15:12:36.643Z
9-updatedAt: 2026-10-04T15:12:50.958Z
9+updatedAt: 2026-10-06T03:59:13.083Z
1010 labels:
1111 - priority::high
12+epic: Agent Memory
1213 body: ""
1314 tasks: []
1415 comments: []
@@ -6,9 +6,10 @@ author:
6 name: k33g6 name: k33g
7 email: ph.charriere@gmail.com7 email: ph.charriere@gmail.com
8 createdAt: 2026-10-04T15:12:36.643Z8 createdAt: 2026-10-04T15:12:36.643Z
9-updatedAt: 2026-10-04T15:12:50.958Z9+updatedAt: 2026-10-06T03:59:13.083Z
10 labels:10 labels:
11 - priority::high11 - priority::high
12+epic: Agent Memory
12 body: ""13 body: ""
13 tasks: []14 tasks: []
14 comments: []15 comments: []
added .tickets/issues/0075-add-gronk-provider.yaml +14 -0
new file mode 100644
@@ -0,0 +1,14 @@
1+# Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
2+id: 75
3+title: Add Gronk provider
4+state: open
5+author:
6+ name: k33g
7+ email: ph.charriere@gmail.com
8+createdAt: 2026-10-06T03:58:37.611Z
9+updatedAt: 2026-10-06T04:00:38.851Z
10+labels: []
11+epic: Providers (LLM Engine)
12+body: ""
13+tasks: []
14+comments: []
new file mode 100644
@@ -0,0 +1,14 @@
1+# Managed by IssueSpec. Hand edits are welcome; keep the schema valid.
2+id: 75
3+title: Add Gronk provider
4+state: open
5+author:
6+ name: k33g
7+ email: ph.charriere@gmail.com
8+createdAt: 2026-10-06T03:58:37.611Z
9+updatedAt: 2026-10-06T04:00:38.851Z
10+labels: []
11+epic: Providers (LLM Engine)
12+body: ""
13+tasks: []
14+comments: []
modified README.md +1 -1
@@ -55,7 +55,7 @@ Type a question at the `>` prompt. Each command the model runs is shown with its
5555 mm -tui
5656 ```
5757
58-The conversation above, a box to type in below. **Enter** sends, **Ctrl-J** adds a line, **Esc** cancels a turn, **F1** lists the keys. With `permissions: ask` in the YAML, the agent asks before running a command or writing a file.
58+The conversation above, a box to type in below. **Enter** sends, **Ctrl-J** adds a line, **Esc** cancels a turn, **F1** lists the keys. With `yolo: false` in the YAML, the agent asks before running a command or writing a file — in the window, the line mode and an editor alike (`yolo: true`, the default, runs them unasked).
5959
6060 ### 3. Inside an editor (ACP)
6161
@@ -55,7 +55,7 @@ Type a question at the `>` prompt. Each command the model runs is shown with its
55 mm -tui55 mm -tui
56 ```56 ```
57 57
58-The conversation above, a box to type in below. **Enter** sends, **Ctrl-J** adds a line, **Esc** cancels a turn, **F1** lists the keys. With `permissions: ask` in the YAML, the agent asks before running a command or writing a file.58+The conversation above, a box to type in below. **Enter** sends, **Ctrl-J** adds a line, **Esc** cancels a turn, **F1** lists the keys. With `yolo: false` in the YAML, the agent asks before running a command or writing a file — in the window, the line mode and an editor alike (`yolo: true`, the default, runs them unasked).
59 59
60 ### 3. Inside an editor (ACP)60 ### 3. Inside an editor (ACP)
61 61
modified agent.dmr.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified agent.llamacpp.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified agent.llmman.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified agent.lsp.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified agent.mcp-example.yaml +5 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified config.samples/agent.dmr.gemma.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified config.samples/agent.dmr.gpt-oss.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified config.samples/agent.dmr.qwen3.6.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified config.samples/agent.llamacpp.ornith.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified config.samples/agent.llamacpp.qwen3.8.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified demo-llmman/agent.jan-nano.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified demo-llmman/agent.llmman.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified demo-llmman/agent.lucy-reviewer.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified demo-llmman/agent.lucy.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified demo-ollama/agent.ollama.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified demo/agent.dmr.2.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified demo/agent.dmr.github-oauth.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified demo/agent.dmr.mcp.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified demo/agent.dmr.sbx.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified demo/agent.dmr.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified demo/agent.lsp.yaml +8 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/diagrams/packages.drawio +2 -2
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/en/README.md +2 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/en/explanation/architecture.md +2 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/en/explanation/asking-another-agent.md +2 -2
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/en/explanation/lsp-tool.md +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/en/explanation/mcp-architecture.md +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
added docs/en/explanation/yolo-and-permissions.md +51 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/en/how-to/ask-another-agent.md +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
added docs/en/how-to/confirm-tool-calls.md +57 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/en/how-to/fetch-a-web-page.md +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/en/how-to/navigate-code-with-lsp.md +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/en/how-to/use-from-an-editor.md +2 -2
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/en/how-to/use-the-window-mode.md +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/en/reference/acp.md +5 -2
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/en/reference/cli.md +14 -5
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/en/reference/configuration.md +5 -3
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/en/reference/mcp-configuration.md +2 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/en/reference/slash-commands.md +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/en/reference/tools.md +4 -2
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/fr/README.md +2 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/fr/explanation/architecture.md +2 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/fr/explanation/asking-another-agent.md +2 -2
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/fr/explanation/lsp-tool.md +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/fr/explanation/mcp-architecture.md +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
added docs/fr/explanation/yolo-and-permissions.md +51 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/fr/how-to/ask-another-agent.md +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
added docs/fr/how-to/confirm-tool-calls.md +57 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/fr/how-to/fetch-a-web-page.md +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/fr/how-to/navigate-code-with-lsp.md +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/fr/how-to/use-from-an-editor.md +2 -2
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/fr/how-to/use-the-window-mode.md +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/fr/reference/acp.md +5 -2
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/fr/reference/cli.md +14 -5
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/fr/reference/configuration.md +5 -3
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/fr/reference/mcp-configuration.md +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/fr/reference/slash-commands.md +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified docs/fr/reference/tools.md +4 -2
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/acp/acp.go +6 -6
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/acp/acp_test.go +14 -6
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/acp/persist.go +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/acp/persist_test.go +2 -2
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/acp/sink.go +33 -18
Large diff collapsed to keep the page fast — load it when you need it.
added internal/acp/sink_test.go +33 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/agent/agent.go +19 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/agent/ask.go +16 -5
Large diff collapsed to keep the page fast — load it when you need it.
added internal/agent/prompt.go +119 -0
Large diff collapsed to keep the page fast — load it when you need it.
added internal/agent/prompt_test.go +101 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/ask/cli.go +11 -3
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/ask/cli_test.go +14 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/ask/client.go +8 -5
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/ask/env.go +6 -4
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/ask/env_test.go +5 -4
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/ask/pool_test.go +28 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/config/config.go +58 -9
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/config/config_test.go +5 -5
Large diff collapsed to keep the page fast — load it when you need it.
added internal/config/yolo_test.go +97 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/mcp/tool_wrapper.go +57 -53
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/tools/fetchcontent.go +7 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/tools/fetchcontent_test.go +1 -0
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/tools/files.go +11 -3
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/tools/lsp.go +5 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/tools/plan.go +27 -16
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/tools/tools.go +5 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/tui/draw.go +2 -2
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/tui/events.go +5 -2
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/tui/sink.go +12 -19
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/tui/turn.go +10 -10
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/tui/turn_test.go +19 -13
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/tui/view.go +10 -7
Large diff collapsed to keep the page fast — load it when you need it.
modified internal/tui/view_test.go +1 -1
Large diff collapsed to keep the page fast — load it when you need it.
modified main.go +23 -13
Large diff collapsed to keep the page fast — load it when you need it.
added main_test.go +22 -0
Large diff collapsed to keep the page fast — load it when you need it.