The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute a variety of local development commands to verify implementations.
Evidence in SKILL.md: Execution of make ci-check for Go backends.
Evidence in SKILL.md: Execution of bun run build, bun run lint, and bun run test for frontend projects.
Evidence in SKILL.md: Execution of docker compose up -d --build and docker compose ps for environment management.
Evidence in SKILL.md: Execution of curl http://localhost:8080/api/health for local connectivity checks.
Evidence in SKILL.md: Execution of a custom review tool via codex review --uncommitted.
[PROMPT_INJECTION]: The skill contains a surface for indirect prompt injection by requiring the agent to follow instructions contained in project-level documentation.
Ingestion points: The agent is instructed to follow procedures in RUNBOOK.md and synchronize state in STATUS.md, SPEC.md, ARCHITECTURE.md, and RUNBOOK.md.
Capability inventory: The agent has the capability to execute shell commands (make, bun, docker, codex) and perform network requests (curl).
Boundary markers: No explicit boundary markers or 'ignore embedded instructions' warnings are provided when processing these files.
Sanitization: No sanitization or validation of the contents of RUNBOOK.md is specified before the agent follows its instructions.

development