The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses child_process.execSync to perform environment discovery and health monitoring. It executes standard commands such as docker --version, git --version, gh auth status, and df -h to verify system tools and resources.
[COMMAND_EXECUTION]: The automated repair functionality in lib/repair.js executes npm install within specific skill directories and performs filesystem modifications (chmod, mkdir) based on diagnostic results.
[EXTERNAL_DOWNLOADS]: The skill performs connectivity tests to well-known and trusted endpoints including Cloudflare (1.1.1.1), GitHub, Anthropic, and OpenAI. These tests use HEAD requests to verify network availability and API responsiveness.
[CREDENTIALS_UNSAFE]: The configuration checks verify the presence of essential environment variables and API keys (e.g., GH_TOKEN, ANTHROPIC_API_KEY, AUTH_SECRET). This logic checks for the existence of the variables without logging or exposing their values.
[PROMPT_INJECTION]: The skill implements a diagnostic surface that processes data from other skill files, creating a potential vector for indirect prompt injection.
Ingestion points: The skill reads SKILL.md files from /job/.pi/skills/* to verify metadata and frontmatter integrity (lib/skills.js).
Boundary markers: Minimal; the check focuses on identifying YAML frontmatter delimiters (---).
Capability inventory: The agent has access to tools that execute subprocesses (execSync), create directories (mkdirSync), and modify permissions (chmodSync) through the diagnostic and repair modules.
Sanitization: Skill names are validated before being used in paths, and repair operations are gated by specific diagnostic findings.

popebot-doctor