The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes standard development tools for verification (type-checking, linting, unit/e2e tests) and version control operations using git and the GitHub CLI (gh). These actions are core to the skill's purpose of managing software releases.\n- [PROMPT_INJECTION]: The skill retrieves build, release, and smoke-check commands from docs/agents/project.md. This represents an indirect prompt injection surface where a compromised project configuration file could influence agent behavior.\n
Ingestion points: Executable commands and release steps are sourced from the local docs/agents/project.md file.\n
Boundary markers: The skill lacks formal delimiters to isolate these external commands, though it explicitly instructs the agent to verify them before execution.\n
Capability inventory: The agent possesses the capability to execute arbitrary shell commands for building/testing and to perform repository changes via git and gh.\n
Sanitization: No programmatic sanitization is applied; the workflow relies on the agent's internal verification and multiple mandatory user approval checkpoints (for versioning, tagging, and publishing) to mitigate risk.

release