fix(lfg): use platform-neutral skill references#642
Merged
Conversation
Replace Claude-specific slash syntax (/ce-plan, /ce-work, etc.) with bare skill names and a platform-agnostic instruction to match the host's available-skills list verbatim. This avoids the agent guessing a short-form name that isn't in the list (e.g., Skill(ce-plan) when the host lists it as compound-engineering:ce-plan) and wasting a turn. Fixes #574
The "orchestration callers pass explicit mode flags" test asserted the literal "/ce-code-review mode:autofix" slash string. Now that lfg uses prose-style "Invoke the `ce-code-review` skill with `mode:autofix ...`", match the intent (ce-code-review invoked with mode:autofix) via regex.
This was referenced Apr 22, 2026
Merged
Closed
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
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Fixes #574. The
lfgskill previously used Claude-specific slash syntax (/ce-plan $ARGUMENTS,/ce-work, etc.) in its step list. When the agent invoked the Skill tool it would try the short name first (e.g.,Skill(ce-plan)), fail because the host listed it ascompound-engineering:ce-plan, then retry with the namespaced form — wasting a turn per step. This PR replaces the slash references with bare skill names and adds a platform-agnostic instruction to match the host's available-skills list verbatim, so the agent resolves the correct entry on the first try across Claude Code, Codex, OpenCode, and other targets.Test plan
/lfg <small task>in Claude Code and confirm the first skill invocation after lfg loads isSkill(compound-engineering:ce-plan, ...)with no prior failedSkill(ce-plan)attempt.