The Agent Skills Directory

[OBFUSCATION]: The file SKILL.md contains the HTML entity < instead of the literal < character within a Bash command example: npm run feature:info -- <feature-id>. While this is likely a rendering artifact for documentation purposes, HTML entities are classified as a form of content obfuscation.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to read and process feature manifest files (YAML format) to validate project health and map features to code, creating a surface for indirect prompt injection.
Ingestion points: The skill reads feature IDs and file paths from external manifest files during npm run feature:info and npm run feature:validate operations (SKILL.md).
Boundary markers: There are no explicit delimiters or instructions provided to the agent to ignore potentially malicious content embedded within the manifest fields.
Capability inventory: The skill has access to Bash (specifically npm and npx commands), Read, Write, and Edit tools as defined in the frontmatter (SKILL.md).
Sanitization: There is no evidence of validation or sanitization of the manifest content before it is processed by the underlying shell tools.

feature-manifest