The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill uses npx to download and execute the @colbymchenry/codegraph package. The installation is performed with the --yes flag, which bypasses the standard user confirmation prompt in the CLI.
[EXTERNAL_DOWNLOADS]: The skill references and fetches content from several external sources, including GitHub repositories for RTK, Serena, and CodeGraph, as well as the Context7 documentation service.
[COMMAND_EXECUTION]: The agent is directed to execute various shell commands to verify tool installation and project status, including git for inspecting changes, curl for fetching web-based documentation, and several tool-specific binaries.
[INDIRECT_PROMPT_INJECTION]: The skill's design requires the agent to ingest data from untrusted local sources (like project Makefiles and package scripts) and external tool help outputs to guide its configuration steps.
Ingestion points: Local configuration files, environment-provided help text, and external documentation.
Boundary markers: Not explicitly defined for data ingestion, but instructions include strict rules against reporting actual credential values.
Capability inventory: File system access, network downloads via curl and npx, and shell command execution.
Sanitization: Explicit instructions to purge credentials and personal session paths from configuration records before committing them to version control.

setup-ai-tooling