feat: import 24 CEP agents and update write-readme command#62
Merged
Conversation
Import 13 new agents and re-sync 11 existing agents from CEP upstream (commit f744b797). All agents processed through mechanical converter pipeline with intelligent rewrite pass for branding, paths, and CC-specific content removal. New agents (13): figma-design-sync, git-history-analyzer, learnings-researcher, repo-research-analyst, agent-native-reviewer, data-integrity-guardian, data-migration-expert, deployment-verification-agent, dhh-rails-reviewer, kieran-rails-reviewer, kieran-typescript-reviewer, lint, pr-comment-resolver Re-synced agents (11): architecture-strategist, best-practices-researcher, bug-reproduction-validator, code-simplicity-reviewer, design-implementation-reviewer, design-iterator, framework-docs-researcher, pattern-recognition-specialist, performance-oracle, security-sentinel, spec-flow-analyzer Intelligent rewrites applied: - repo-research-analyst: **/CLAUDE.md glob -> **/AGENTS.md - learnings-researcher: fixed relative path to compound-docs skill - git-history-analyzer: compound-engineering -> systematic branding - best-practices-researcher: .claude/skills/ -> .opencode/skills/ - code-simplicity-reviewer: removed CC pipeline artifacts exclusion - pattern-recognition-specialist: grep -> Grep tool naming Command fixes: - lfg.md: removed CC-specific commands, fixed prefixes - deepen-plan.md: replaced all .claude/ paths, fixed tool references - agent-native-audit.md: fixed /compound-engineering: prefix - review.md: removed 3 phantom agents, added kieran-typescript-reviewer Manifest: 29 tracked definitions (24 agents + 5 commands) Verified: build, typecheck, lint (0 issues), tests (328/328 pass)
Address 6 gaps discovered during the 24-agent import batch: 1. Fix upstream_path in manifest example to use full plugins/compound-engineering/ prefix 2. Document both plain string and structured object formats for manual_overrides (plain strings for initial import) 3. Add phantom agent check to evaluation criteria — commands can reference agents that don't exist upstream 4. Document color frontmatter field passthrough behavior 5. Add skill path patterns (~/.claude/skills/, .claude/skills/) and CC-specific exclusion rules to converter gap list 6. Add batch fetch pattern and shared-metadata step to Batch Import section Also adds tool name capitalization and compound-engineering pipeline artifacts to branding audit table.
Rewrite write-readme command to use live CLI inventory, @file references, and formatting rules matching the evolved README style. Remove stale template and fragile find-exec shell injections. Update README.md agent tables from 11 to 24 agents across all 4 categories (design, research, review, workflow). Fix counts in Key Features and project structure tree. Add manifest.ts to structure.
Add comprehensive solution documentation capturing the process, challenges, and learnings from importing 24 agents from CEP upstream. Key content: - Problem analysis: phantom agent references, missing imports, stale docs - Investigation: inventory of upstream agents, relevance assessment - Root cause: partial migration with no manifest tracking - Solution: batch import workflow with convert-cc-defs skill - Rewrites: 6 agents required domain-specific adaptations (paths, branding) - Command fixes: lfg, deepen-plan, agent-native-audit, review - Manifest structure: upstream tracking with content hashes for drift detection - Prevention: automated checks, phantom agent detection patterns - Verification: all tests passing, 24 agents imported, 29 manifest entries This documents the systematic approach to importing definitions from external sources while maintaining provenance and preventing stale references. Related: convert-cc-defs skill, sync-manifest.json, 24 new agents
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
f744b797through the full convert-cc-defs workflow (mechanical conversion + intelligent rewrite + manifest tracking)bun src/cli.ts list),@filereferences, and formatting rules matching the evolved README styleNew Agents (13)
figma-design-syncgit-history-analyzerlearnings-researcherrepo-research-analystagent-native-reviewerdata-integrity-guardiandata-migration-expertdeployment-verification-agentdhh-rails-reviewerkieran-rails-reviewerkieran-typescript-reviewerlintpr-comment-resolverIntelligent Rewrites Applied
CLAUDE.mdglobs →AGENTS.md.claude/skills/paths →.opencode/skills/and~/.config/opencode/skills/compound-engineeringbranding →systematicgreptool →Greptool namingcommands/workflows/review.mdCommand Fixes
commands/lfg.md: removed CC-specific commands, fixedsystematic:prefixcommands/deepen-plan.md: replaced all.claude/paths, fixed tool referencescommands/agent-native-audit.md: fixed/compound-engineering:→/systematic:commands/workflows/review.md: removed 3 phantom agents, added 2 real agentsSkipped (org-specific)
every-style-editor,ankane-readme-writer,kieran-python-reviewer,julik-frontend-races-reviewer,schema-drift-detectorVerification