The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through user-controlled parameters that are used as arguments for deployment scripts.
Ingestion points: Arguments such as --version, --url, and --db-url passed to scripts/deploy.sh, scripts/smoke-test.sh, and scripts/migration-dry-run.sh respectively.
Boundary markers: Absent. The skill does not implement specific delimiters or instructions to ignore embedded commands within these input strings.
Capability inventory: The skill utilizes the Bash tool to perform Docker operations, GitHub CLI tasks, and Git commands. It also executes database operations via psql and alembic.
Sanitization: While the scripts utilize standard shell quoting for variables, they lack advanced sanitization or validation for identifiers used in SQL commands or complex shell strings, which could lead to injection under specific adversarial conditions.
[COMMAND_EXECUTION]: The skill uses the Bash tool to orchestrate deployment tasks including Docker service updates, Git repository management, and environment validation.
[EXTERNAL_DOWNLOADS]: The skill fetches various software dependencies and automation templates from well-known public registries.
Python dependencies: Installs packages such as ruff, mypy, pytest, and alembic from PyPI.
Node.js dependencies: Installs packages including eslint, prettier, and playwright from NPM.
GitHub Actions: The CI/CD template utilizes official actions from the actions organization (e.g., actions/checkout@v4, actions/setup-python@v5) for pipeline automation.

deployment-pipeline