The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses standard shell commands like ls and date for file discovery and timestamp calculations. It also utilizes the GitHub CLI (gh) to retrieve repository events, commits, and pull requests. These actions are aligned with the skill's purpose for developers.
[EXTERNAL_DOWNLOADS]: The push and scorecard branches interact with the official GitHub API to fetch repository data such as events and commit diffs. These operations are limited to metadata retrieval for reporting purposes and target standard development infrastructure.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data (commit messages, PR titles, and diffs) which represents an attack surface for indirect prompt injection where malicious content in a repository could influence the recap's synthesis. Ingestion points: Commit messages, PR titles, and diffs fetched via gh api, along with local logs in memory/logs/. Boundary markers: The instructions do not define specific delimiters or instructions to ignore embedded prompts within the fetched data. Capability inventory: Writing report files to output/articles/ and apps/dashboard/, and sending messages via the internal ./notify tool. Sanitization: The skill does not explicitly describe sanitization or escaping of the external content before inclusion in the final reports.

operator-scorecard