The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes a local Python script located at .codestable/tools/search-yaml.py. This script is used to query metadata within the project's documentation directories (docs/dev and docs/user) to manage document versions and statuses.
[DATA_EXPOSURE]: The instructions direct the agent to read internal project files, such as .codestable/attention.md and various specification documents, to gather the necessary context for generating accurate manuals. This access is restricted to the local project environment and is essential for the skill's functionality.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it ingests untrusted data from project source code and external specification documents during the documentation drafting process.
Ingestion points: Reads content from the docs/ directory, project source code, and internal specification documents (SKILL.md).
Boundary markers: The instructions do not specify any delimiters or safety markers to differentiate between documentation content and potential embedded instructions.
Capability inventory: The agent can search the filesystem using a local script and write new markdown files to the docs/ directory.
Sanitization: No specific sanitization or validation logic is defined for the external content processed by the skill.

cs-doc-tutorial