The Agent Skills Directory

[SAFE]: The skill performs legitimate content generation by reading local project documentation (PRD, README) and producing social media drafts. No malicious patterns were identified.
[COMMAND_EXECUTION]: The skill executes git log to retrieve recent commit messages for release notes. The command is hardcoded and does not include unsanitized user input, making it safe for use in development environments.
[DATA_EXFILTRATION]: No network-capable tools are requested in the skill's configuration (allowed-tools). Data is processed locally and written to the project's own docs/ directory.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted content from project files (PRDs, READMEs).
Ingestion points: Reads content from docs/prd.md, README.md, CLAUDE.md, and research.md.
Boundary markers: No specific delimiters or boundary markers are used when interpolating file content into prompts.
Capability inventory: Has permissions to write to the filesystem (Write) and execute a specific git command.
Sanitization: No explicit sanitization or filtering is performed on the ingested content. However, the risk is low as there is no network access and the output is directed to a local documentation file for human review.

solo-content-gen