docs: the autonomy contract — AUTONOMY.md, model-tier doctrine#35
Merged
Conversation
Adds AUTONOMY.md, the canonical page mapping each Mind-prompt Autonomy: level (safe/supervised/human-required) to behaviour at every human checkpoint, with per-work-type caps, the explicit --auto activation rule, the calibration-log spec, and hard invariants. WORKFLOW.md links the contract and its model-delegation doctrine is now tier-based (judgment tier currently Fable 5; execution tier Sonnet). AGENTS.md gains a pointer. Doctrine only — no skill behaviour changes; task 1 of PyAutoMind/feature/autonomy/. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_013UptUDNTYiB7izaQJBAKgS
There was a problem hiding this comment.
Pull request overview
Adds the canonical autonomy contract documentation for PyAutoBrain, defining how the Autonomy: header levels map onto human checkpoints in the dev workflow, and updates the workflow doctrine to be model-tier based (so it stays stable as model availability changes).
Changes:
- Add
AUTONOMY.mdas the single-source “autonomy contract” (checkpoints, per-level behavior, caps, activation rule, invariants). - Update
skills/WORKFLOW.mdto link autonomy checkpoint rules to the contract and to describe model delegation in terms of judgment/execution tiers. - Add a pointer in
AGENTS.mdto the new autonomy contract page.
Reviewed changes
Copilot reviewed 3 out of 3 changed files in this pull request and generated 1 comment.
| File | Description |
|---|---|
skills/WORKFLOW.md |
Introduces an Autonomy section linking to the contract; updates model delegation doctrine to tier-based terminology. |
AUTONOMY.md |
New canonical autonomy contract defining checkpoints, per-level behavior, caps, activation, and invariants. |
AGENTS.md |
Adds a single pointer line to AUTONOMY.md to avoid duplicated autonomy prose. |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Comment on lines
+76
to
+77
| The workflow skills split work across **model tiers**, not named models — the | ||
| doctrine survives model access changing: |
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Adds
AUTONOMY.md, the one canonical page defining what each Mind-promptAutonomy:level (safe/supervised/human-required) does at each of the six human checkpoints (plan approval, ship PR sign-off, Heart YELLOW acknowledgement, merge/close, pre_build version ask, post-merge cleanup); per-work-type caps (refactor/test/maintenance may runsafe, feature/bug/docs cappedsupervised, release alwayshuman-required); the explicit---autoactivation rule; the calibration-log spec (PyAutoMind/autonomy_log.md); and the hard invariants (merge always human, autonomous runs end at PR-open, Heart YELLOW never acknowledged autonomously).skills/WORKFLOW.mdgains an "Autonomy" section linking the contract, and its model-delegation doctrine becomes tier-based — judgment tier = strongest available model (currently Fable 5, fallback Opus); execution tier = fast/cheap (currently Sonnet) — so the doctrine survives model access changing.AGENTS.mdgains a pointer line.Doctrine only — no skill behaviour changes; consumption lands with
PyAutoMind/feature/autonomy/tasks 2–7 (review faculty, ship gate,--automode, checkpoint-and-continue, queue runner). Task 1 of the autonomy series.Closes #34.
API Changes
None — documentation/doctrine only.
Test Plan
AUTONOMY.mdlinks resolve fromAGENTS.mdandskills/WORKFLOW.md--autolaunchFull API Changes (for automation & release notes)
None — internal documentation/doctrine only.
🤖 Generated with Claude Code
https://claude.ai/code/session_013UptUDNTYiB7izaQJBAKgS