fix(skills): descriptions fit the hosts' skills context budget - #31
Merged
Merged
Conversation
Every SKILL.md description is now the skill's resource.yaml summary
(<=160 chars): 94 skills go from 31,725 to 13,322 characters of
description injected into every session. Codex had started truncating
("Skill descriptions were shortened to fit the skills context budget")
and Claude Code's budget is about 15k characters for all skills. Trigger
phrases stay in the skill bodies; the routing eval reads summaries and
shows no regression. validate-library now fails when a description and
its summary differ, and the manifests test enforces the 160-char line.
Co-Authored-By: Claude Fable 5.1 <noreply@anthropic.com>
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Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be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
Codex reported: "Skill descriptions were shortened to fit the skills context budget." Our 94 SKILL.md descriptions totalled 31,725 characters (about 7,900 tokens) injected into every session, each carrying a long "Trigger on ..." list. Claude Code's budget for all skill descriptions is about 15k characters, so it was likely truncating silently too.
descriptionis now the skill'sresource.yamlsummary(≤160 chars): 13,322 characters total, about 3,300 tokens.validate-libraryfails when the two differ; the manifests test enforces the one-line budget.npm run eval:routingreads summaries and shows no regression.docs/SKILLS.md; CHANGELOG under 3.1.1 (not yet published).Test plan
npx vitest run1483 pass,npm run validate:library,npm run generate:registry -- --check,npm run eval:routing(no regression)🤖 Generated with Claude Code