feat(ce-work): accept bare prompts and add test discovery#423
Merged
Conversation
…n loop Add Phase 0 input triage so ce:work can be invoked with a raw prompt instead of requiring a plan file. Complexity assessment routes trivial work directly to execution, nudges large work toward brainstorm/plan without forcing it, and builds inline task lists for everything in between. Add Test Discovery as a universal execution step — find existing test files for implementation files before changing them, and ensure test updates (add/modify/remove) accompany implementation changes regardless of whether a plan enumerated them. Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>
Sync decision: Propagated to beta — Phase 0 input triage and Test Discovery are general execution improvements orthogonal to the delegate mode that beta exists to test. Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: f80aa043ae
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
Trivial path now says "implement directly — no task list, no execution loop" instead of "skip task tracking" which contradicted the Phase 2 while-loop. Strategy selection now defaults bare prompts to inline and marks subagent strategies as requiring plan-unit metadata. Both fixes applied to ce-work and ce-work-beta. Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 96856e7896
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
The skip condition on "Create Todo List" only covered "Phase 0 already built one" but trivial work intentionally builds no task list. Extended the skip to also cover the trivial route. Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>
Merged
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
1 participant
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
ce:work now handles bare prompts (not just plan files) and includes test discovery as a first-class execution step.
Phase 0 — Input Triage routes bare prompts through a complexity assessment before execution. Trivial work skips ceremony entirely, small/medium work gets an inline task list, and large work surfaces a recommendation toward brainstorm/plan without forcing it. When a plan document is provided, Phase 0 is skipped entirely — no overhead for the existing pipeline.
Test Discovery is a new universal step in the Phase 2 execution loop. Before changing a file, the agent finds its existing test files (by import, naming pattern, or shared references). When a plan specifies test scenarios, the agent starts there and supplements with any coverage the plan missed. Test guidance throughout the skill now uses "add/update/remove" instead of just "add new tests."
Both changes propagated to ce:work-beta — they're orthogonal to the delegate mode that beta exists to test.
🤖 Generated with Claude Opus 4.6 (1M context) via Claude Code