feat: add 156 specialist agent catalog with setup integration and /design skill enhancement#34
Open
jakub-kaspar-git wants to merge 4 commits intostranma:masterfrom
Open
Conversation
Add scripts to convert agency-agents Markdown files to the template's structured agent format and validate the output. - scripts/convert_agents.py: Bulk-converts agents with frontmatter mapping, emoji stripping, section filtering, model/tools/permission heuristics - scripts/validate_agents.py: Validates converted agents have correct YAML frontmatter with required fields and valid values - tests/test_agent_catalog.py: 43 tests covering parsing, slugification, body transformation, conversion, and validation Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>
Auto-generated via `scripts/convert_agents.py` from the agency-agents repository (https://github.com/webdevtodayjason/agency-agents). 156 agents across 14 categories: - academic (5), design (8), engineering (23), game-development (20) - marketing (27), paid-media (7), product (5), project-management (6) - sales (8), spatial-computing (6), specialized (27) - support (6), testing (8) Each agent is converted to the template format with: - Structured YAML frontmatter (model, tools, permissionMode, color) - Action-oriented body (personality sections stripped/condensed) - Category-prefixed kebab-case naming Includes manifest.json for programmatic catalog access. Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>
Extend the setup script so users can select which agent categories to install during project initialization. Changes: - setup_project.py: Add --agents flag (all/none/comma-separated), --keep-catalog flag, interactive category selector, install and cleanup functions - README.md: Document Agent Catalog feature with usage examples and category table - tests/test_agent_catalog.py: Add setup integration tests (install, cleanup, list categories) Usage: python setup_project.py --name my-project --agents all python setup_project.py --name my-project --agents "engineering,testing" Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>
Add Step 0 to /design that optionally loads a specialist agent from the agent catalog for domain-specific planning expertise. Usage: /design build login page using engineering-frontend-developer /design optimize queries using engineering-database-optimizer When an agent is loaded, its Core Mission, Critical Rules, and Workflow Process are incorporated into the plan's approach, risks, and priorities. When no agent is specified, suggests relevant specialists from the catalog. Co-Authored-By: Claude Opus 4.6 (1M context) <[email protected]>
|
Important Review skippedToo many files! This PR contains 163 files, which is 13 over the limit of 150. ⚙️ Run configurationConfiguration used: Organization UI Review profile: ASSERTIVE Plan: Pro Run ID: 📒 Files selected for processing (163)
You can disable this status message by setting the Use the checkbox below for a quick retry:
✨ Finishing Touches🧪 Generate unit tests (beta)
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. Comment |
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
setup_project.pywith--agentsflag for category-based agent selection (interactive + CLI)/designskill to optionally load specialist agents for domain-specific planningMotivation
The template ships with 6 built-in workflow agents (code-quality, test-coverage, pr-writer, etc.) that handle the development process. However, users working in specific domains (frontend, DevOps, security, game dev, marketing, etc.) lack access to domain-specific expertise during planning and implementation.
The agency-agents repository provides 152+ production-ready AI agent personalities covering 14 professional domains. This PR integrates them into the template as an optional, selectable catalog -- users pick what they need during
setup_project.py, keeping the default template lean.Architecture
Key design decisions:
.claude/agent-catalog/which is NOT scanned by Claude Code (only.claude/agents/is).claude/agents/during setup, then the catalog is deletedmodel,tools,permissionMode) -- not used as-is from sourcescripts/convert_agents.pyagainst the source repoChanges
Features
Commit 1 -- Conversion tooling (
scripts/convert_agents.py,scripts/validate_agents.py)convert_agents.py: Bulk-converts agency-agents Markdown to template formatname-> kebab-case slug, addsmodel/tools/permissionMode)acceptEdits, testing ->dontAsk)manifest.jsoncatalog indexvalidate_agents.py: Validates all converted agents have correct frontmatter and structureCommit 2 -- Agent catalog (
.claude/agent-catalog/)156 converted agents across 14 categories:
All auto-generated by
scripts/convert_agents.py-- reviewer should spot-check 2-3 agents per category rather than reviewing all 156Commit 3 -- Setup integration (
setup_project.py,README.md)--agentsCLI flag:all,none, or comma-separated categories--keep-catalogflag to preserve catalog after setupinstall_agent_catalog(),cleanup_agent_catalog(),list_agent_categories()Commit 4 -- /design skill enhancement (
.claude/skills/design/SKILL.md)using <agent-name>from input.claude/agents/then.claude/agent-catalog/**/for the agent file/design build responsive dashboard using engineering-frontend-developerTests
Documentation
README.md: New "Agent Catalog" section with usage examples, category table, and--agents/--keep-catalogflagsTest Plan
uv run pytest)python -m pytest tests/test_agent_catalog.py -v)python scripts/validate_agents.py .claude/agent-catalog/)setup_project.py --name test --agents engineering,testing --keep-setup --keep-catalogin temp dir, verify agents land in.claude/agents//design build API using engineering-backend-architectloads agent contextAcceptance Criteria
setup_project.py --agents allinstalls all agents to.claude/agents/setup_project.py --agents none(default) changes nothing -- template stays lean--keep-catalog)--agentsflag/designskill loads specialist agent whenusing <name>is provided/designsuggests relevant agents when no specialist specifiedscripts/convert_agents.pycan regenerate the catalogReviewer Guide
This PR adds 163 files / ~35K lines, but most of it is auto-generated agent content. Here's how to review efficiently:
Review carefully (human-written code, ~1500 lines):
scripts/convert_agents.py-- conversion logic and heuristicsscripts/validate_agents.py-- validation rulessetup_project.pydiff -- new functions and CLI integrationtests/test_agent_catalog.py-- test coverage.claude/skills/design/SKILL.mddiff -- skill enhancementREADME.mddiff -- documentationSpot-check (auto-generated, ~34K lines):
.claude/agent-catalog/name,model,tools,permissionMode,colormanifest.json-- verify category counts match directory contentsSkip (verified by automated tests):
test_all_agents_validGenerated with Claude Code