The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes several command-line tools including gh (GitHub CLI), jq, python3, awk, and grep to interact with GitHub Projects and process data.
Evidence found in references/board-queries.md, which contains bash scripts for board resolution, item enumeration, and mutation.
Evidence in the enumeration step using a python3 one-liner to validate JSON data lengths.
[INDIRECT_PROMPT_INJECTION]: The skill reads and analyzes GitHub issue bodies, which are untrusted external inputs, to drive decision-making for board updates.
Ingestion points: The content.body of issues is retrieved in references/board-queries.md for triage analysis.
Boundary markers: The instructions in references/triage-rubric.md do not specify delimiters or warnings to ignore malicious instructions embedded in issue bodies.
Capability inventory: The skill is capable of performing data mutations on GitHub Projects using gh api graphql and modifying labels via gh issue edit as described in references/board-queries.md.
Sanitization: No explicit sanitization or filtering of issue content is performed prior to the agent's evaluation against the triage rubric.

triage-board