agentleFS
Sign inStart free
CrowdStrikeCrowdStrike/agentworks-skills/CLAUDE.md

agentworks-skills

CLAUDE.md2 starsChanged 22 days agoMITView on GitHub

Project-level instructions for Claude Code when working in the agentworks-skills plugin. For the tool-agnostic guide (repo structure, skills ecosystem, usage without the plugin system), see AGENTS.md. This plugin includes two hooks that run automatically: Both hooks are advisory only — they always exit 0 and never block a user action or a tool call. They need jq and stay silent without it. Each thought on the left has led to a wrong result; the right column says what to do…

The file

We split this CLAUDE.md into the rules it gives an agent

2 rules. About 0 are common, 1 is distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Shared auth: All Python scripts import credentials and FalconPy clients from common/scripts/auth.py (get_agents_client(), …

    line 32
  2. Distinctive:

    Superpowers: If installed, superpowers planning/TDD skills MAY supplement the workflow but should not replace the orchestrator's routing.

    line 34

≈ “Common” is an estimate: at least 1 in 5 of the 26,980 CLAUDE.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
CLAUDE.md

CLAUDE.md

Project-level instructions for Claude Code when working in the agentworks-skills plugin. For the tool-agnostic guide (repo structure, skills ecosystem, usage without the plugin system), see AGENTS.md.

Plugin Hook Behavior

This plugin includes two hooks that run automatically:

  • SessionStart: bootstrap.sh builds or refreshes the managed Python venv. It is a no-op when requirements.txt and the venv's Python are unchanged, and it never blocks the session.
  • UserPromptSubmit + PreToolUse (Bash|Skill): agentworks-router.sh classifies only the user's prompt. On a match it writes a session-scoped marker and injects advisory context pointing at the agentworks orchestrator skill. On the next Bash call it nudges once for that prompt; loading any skill clears the marker.
  • PreToolUse (Skill): agentworks-foundry-bridge.sh gives advisory cross-plugin routing between this plugin (API-managed agents and knowledge bases) and crowdstrike-falcon-foundry (agents defined in a Foundry app's manifest.yml). It stays silent unless the prompt was about one of the two.

Both hooks are advisory only — they always exit 0 and never block a user action or a tool call. They need jq and stay silent without it.

Counter-Rationalizations

Each thought on the left has led to a wrong result; the right column says what to do instead:

ThoughtReality
"I'll invoke the draft with invoke_agent.py"invoke_agent.py only runs the published version. Use invoke_version.py --version-id for drafts
"I know the agent ID"Query first (agent_search.py); never guess IDs
"I'll suggest prompt changes from the description"Analyze real traces first (analyze_agent.py, inspect_invocation.py)
"I'll delete or disable the agent"Delete and enable/disable are out of scope for v1
"I'll call the API directly"Use the scripts; they handle auth, pagination, and HTTP errors

Skills Integration

  • Orchestration: The agentworks skill is the entry point. It routes to agents, invocation, knowledge-bases, discovery, and setup.
  • Direct invocation: Each sub-skill can be used directly for a focused task.
  • Shared auth: All Python scripts import credentials and FalconPy clients from common/scripts/auth.py (get_agents_client(), get_agent_invocation_client(), get_kb_client(), …), which share one OAuth2 token.
  • Shared helpers: common/scripts/formatting.py (durations, timestamps), discovery_helpers.py (span queries, hydration), invocation_helpers.py (invocation status constants), and _bootstrap.py (re-exec through the managed venv).
  • Superpowers: If installed, superpowers planning/TDD skills MAY supplement the workflow but should not replace the orchestrator's routing.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free