The Agent Skills Directory

[SAFE]: The skill operates as an advisory tool for developer relations. It provides templates for releases, announcements, and tracking issues, along with diagnostic logic based on repository metrics. No evidence of malicious intent, obfuscation, or safety bypass was found.
[INDIRECT_PROMPT_INJECTION]: The skill has an attack surface for indirect prompt injection as it processes data from external sources and user-provided content.
Ingestion points: Data retrieved from GitHub API endpoints (traffic, views, clones), web scraping of the public GitHub 'Used by' page in scripts/github-dependents.gs, and agent requests for user-provided pastes of repository content in SKILL.md.
Boundary markers: None. The skill does not specify any delimiters or instructions to the agent to isolate external data from its primary instructions.
Capability inventory: The agent's capabilities are limited to diagnostic classification, template population, and plan formulation. The provided scripts perform network requests to official GitHub domains for data collection.
Sanitization: There is no explicit sanitization or filtering logic implemented for data ingested from GitHub or user pastes.
[DATA_EXPOSURE_AND_EXFILTRATION]: The included Google Apps Scripts are used to archive repository engagement data. These scripts follow secure practices by retrieving authentication tokens from Google Apps Script properties (ScriptProperties) rather than hardcoding them and communicate exclusively with well-known GitHub domains (api.github.com and github.com).

github-engagement-playbook