The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data from multiple external sources, which could contain malicious instructions designed to influence the agent's behavior.
Ingestion points: The skill reads all in-repo documentation (Phase 1a in SKILL.md), narrative guides, migration guides, and GitHub issues/discussions (Phase 3 in references/deep-read.md).
Boundary markers: The instructions do not specify the use of delimiters or specific instructions to ignore embedded prompts within the documentation or issue content.
Capability inventory: The agent has the capability to read local files and perform network operations via the gh tool (gh search issues, gh search prs, gh api).
Sanitization: There are no explicit steps provided for sanitizing or filtering instructions that might be embedded in the fetched documentation or GitHub metadata.
[COMMAND_EXECUTION]: The skill explicitly instructs the agent to execute shell commands to interact with GitHub services.
Evidence: references/deep-read.md directs the agent to use gh search issues, gh search prs, and gh api to extract failure modes and developer confusion patterns.

domain-discovery