The Agent Skills Directory

[COMMAND_EXECUTION]: Executes a bundled bash script scripts/get-standup-context.sh and uses the gh (GitHub CLI) to perform API queries and mutations.
[DATA_EXFILTRATION]: Collects PR, issue, and discussion metadata from specified GitHub organizations and posts it to a discussion in the mitodl/hq repository. This behavior is consistent with the skill's documented purpose of automating standup reports for the author's organization.
[INDIRECT_PROMPT_INJECTION]: The skill processes data from GitHub PR and issue titles which could contain malicious strings designed to influence agent behavior.
Ingestion points: Fetches data via gh search in scripts/get-standup-context.sh.
Boundary markers: Absent; data is processed and rendered into a Markdown template without specific markers to isolate untrusted content.
Capability inventory: Includes shell command execution and write access to GitHub repository discussions via the GraphQL API.
Sanitization: Uses jq to safely escape variables interpolated into shell commands, preventing command injection, but does not perform semantic sanitization of the fetched text content.

generate-standup