Conversation
Member
|
Closing. how-to-flow is fully autonomous by design; pausing a pivot for a user decision while "continuing work on the original goal" is not coherent inside an unattended loop, and relaxing the preflight fact-loading gate removes what makes the loop CDCL-like. The concern (drift reported as solved) is already covered by the final check on main, which requires the original goal, not a drifted one, and PIVOTED-SOLVED already reports the gap to the original. |
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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.
The autonomous flow workflow could relax a hard goal during a pivot and still report the result as solved. This PR keeps the original goal and test authoritative: a materially easier target becomes a proposal for the user, and accepted changes are recorded as a new goal rather than silently weakening success criteria.
The change follows OpenAI's Rethinking skills and prompts for GPT-6 Astra, especially explicit authorization boundaries and honest completion states. It is the flow slice split from #62 after review feedback.
Validation
python3 scripts/validate_skills.py: 16 skills passpython3 -m pytest -q: 268 passed, 5 skippedgit diff --check: clean