t258: Exclude skill reference files from subagent stub generation

[marcusquinn]

Summary

• Add -not -name '*-skill.md' filter to find command in generate-opencode-agents.sh
• Prevents skill reference files from being processed as OpenCode subagents
• Skill files are documentation/reference only, not executable agents

Changes

• Modified .agents/scripts/generate-opencode-agents.sh line 782
• Added filter alongside existing -not -path '*/loop-state/*' exclusion

Testing

• Run generate-opencode-agents.sh
• Verify skill files are excluded from subagent generation (12 skill files excluded, 893 subagents processed)
• Run shellcheck validation (passes with no violations)

Verification

# Before: Would process all .md files including skills
find ~/.aidevops/agents -mindepth 2 -name '*.md' -type f -not -path '*/loop-state/*' | wc -l
# 905 files

# After: Excludes skill files
find ~/.aidevops/agents -mindepth 2 -name '*.md' -type f -not -path '*/loop-state/*' -not -name '*-skill.md' | wc -l
# 893 files (12 skill files excluded)

Related: t258

Summary by CodeRabbit

• Chores
  • Updated internal build script to refine file discovery and processing logic.

[gemini-code-assist]

Warning

You have reached your daily quota limit. Please wait up to 24 hours and I will start processing your requests again!

[coderabbitai]

Caution

Review failed

The pull request is closed.

Walkthrough

A shell script modification that adds an exclusion filter to the subagent markdown generation process, preventing files matching "*-skill.md" from being processed by the find command while preserving all downstream logic and workflows.

Changes

Cohort / File(s)	Summary
Agent Generation Script .agents/scripts/generate-opencode-agents.sh	Added -not -name "*-skill.md" exclusion pattern to the find command, preventing skill markdown files from triggering subagent stub generation.

Estimated code review effort

🎯 1 (Trivial) | ⏱️ ~3 minutes

Possibly related issues

• t258: Exclude skill reference files from subagent stub generation — references/ dirs inside imported skills (mermaid-diagrams, modern-javascript, postgres-drizzle, etc.) generate 30+ stubs with generic names (SCHEMA, PROMISES, api, configuration) that pollute the subagent namespace and risk name collisions across skills. Fix: add -not -path "*/references/*" to the find command in generate-opencode-agents.sh. Reference files are internal to the parent skill and should not be invocable as standalone subagents. Also consider excluding templates/plugin-template/ for the same reason. #1026 — Addresses file exclusion patterns in the same generation script for filtering subagent discovery.

Possibly related PRs

• fix(opencode): exclude loop-state from agent discovery and deployment #110 — Modifies the same find command in generate-opencode-agents.sh to add additional exclusion patterns for structured file organization.
• feat: use source file description in generated subagent stubs (t255) #1017 — Updates the subagent stub generation logic in the same script to refine how agent discovery operates.
• chore: complete t062 research on vercel-labs/agent-skills #559 — Documents SKILL.md generation and adoption patterns, contextualizing why skill files require exclusion handling.

Poem

🎯 Filter flows with measured grace,
Skill files find their rightful place,
One small flag, one clean exclusion,
No more stubs from the right solution! ✨

✨ Finishing touches

• 📝 Generate docstrings

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Post copyable unit tests in a comment
• Commit unit tests in branch feature/t258

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands and usage tips.}

[github-actions]

🔍 Code Quality Report

�[0;35m[MONITOR]�[0m Code Review Monitoring Report

�[0;34m[INFO]�[0m Latest Quality Status:
SonarCloud: 0 bugs, 0 vulnerabilities, 46 code smells

�[0;34m[INFO]�[0m Recent monitoring activity:
Wed Feb 11 00:30:54 UTC 2026: Code review monitoring started
Wed Feb 11 00:30:55 UTC 2026: SonarCloud - Bugs: 0, Vulnerabilities: 0, Code Smells: 46

📈 Current Quality Metrics

• BUGS: 0
• CODE SMELLS: 46
• VULNERABILITIES: 0

Generated on: Wed Feb 11 00:30:57 UTC 2026

---

Generated by AI DevOps Framework Code Review Monitoring

[gemini-code-assist]

Warning

You have reached your daily quota limit. Please wait up to 24 hours and I will start processing your requests again!

[sonarqubecloud]

Quality Gate passed

Issues
0 New issues
0 Accepted issues

Measures
0 Security Hotspots
0.0% Coverage on New Code
0.0% Duplication on New Code

See analysis details on SonarQube Cloud