The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process a wide variety of untrusted data from the project environment to generate documentation artifacts.
Ingestion points: The documentation generation modes (api, data, e2e, test-plan) read source code files, OpenAPI specifications, database schemas, and existing Jira issue descriptions to extract business logic and requirements.
Boundary markers: The skill consistently requires a 'diff summary' and explicit user approval before writing or updating any artifacts, which serves as a critical human-in-the-loop security control.
Capability inventory: The skill has access to database tools, API schema processing, diagram generation, and can perform writes to local files and the project's Jira instance.
Sanitization: While it uses specialized scripts to convert content formats (e.g., markdown to Jira ADF), it lacks explicit sanitization for instructions that might be embedded within the processed source code (such as malicious comments intended to influence the agent's summary).
[COMMAND_EXECUTION]: The skill relies on executing local project-specific scripts and management tools to perform its discovery and synchronization tasks.
Evidence: The instructions frequently invoke shell commands such as bun run context:map, bun run api:sync, and bun run jira:sync-issues to interact with the project's tooling ecosystem.
These executions are central to the skill's functionality and assume a trusted local environment where these scripts are defined.

project-context