The Agent Skills Directory

[DYNAMIC_CONTEXT_INJECTION]: The skill utilizes exclamation mark and backtick syntax placeholders to execute shell commands at load time for document validation. These commands, including grep and test, are used to verify the existence of files and the integrity of Markdown table structures within the local repository. These operations are limited to local repository inspection and do not involve network access or sensitive file exposure.
[INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface as it ingests data from external SKILL.md files to generate documentation indices.
Ingestion points: Scanning skills/*/SKILL.md files to extract frontmatter data.
Boundary markers: None explicitly defined for preventing the interpretation of extracted data as instructions.
Capability inventory: File read/write operations and shell command execution (grep, test) in SKILL.md patterns.
Sanitization: No specific sanitization or escaping logic is described for the extracted metadata before it is written to index files.

documentation-update