The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to discover repository structure, inventory existing scripts, and execute the generated development CLI or just commands. These operations are restricted to the local environment and are necessary for the skill's primary function of repository reconciliation.
[DATA_EXPOSURE]: The agent reads sensitive configuration files such as .env and .env.example to inventory and migrate environment variables. It also accesses repository metadata (remotes) and interacts with issue trackers (GitHub, Jira) to establish project bindings. The instructions include specific logic to preserve secrets during these operations.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from existing repository scripts and issue tracker content (titles, bodies, labels). This represents an attack surface for indirect prompt injection.
Ingestion points: Local project scripts (scripts/*.mjs, etc.), configuration files, and external issue tracker data via CLI tools (gh, jira).
Boundary markers: None identified in the prompt templates or instructions.
Capability inventory: Access to Bash, Write, Edit, and Read tools.
Sanitization: No explicit data validation or sanitization is performed on ingested external content.

setup-engineer