The Agent Skills Directory

[SAFE]: The skill initializes project documentation and setup scripts locally. It performs standard file operations (create, copy, write) and runs provided Node.js utilities for verification. All operations are confined to the workspace and no external network activity or credential access was observed.
[INDIRECT_PROMPT_INJECTION]: The skill possesses a potential surface for indirect prompt injection as it ingests untrusted project data (such as README files, lockfiles, and directory structures) to classify the project and generate documentation drafts.
Ingestion points: The skill reads local repository files including README.md, build configurations, and directory trees as described in references/classify.md and references/interview.md.
Boundary markers: The instructions do not specify explicit delimiters or warnings to ignore embedded instructions within the ingested project files.
Capability inventory: The skill is capable of performing file system writes (creating and overwriting documentation files) and executing local Node.js scripts (precheck.mjs) for validation.
Sanitization: No explicit sanitization or filtering of text extracted from existing project documentation is mentioned; the agent is instructed to summarize and categorize the findings.

setup