The Agent Skills Directory

[PROMPT_INJECTION]: The skill ingests and parses project files (.tex and .bib) to extract citation keys. This creates an indirect prompt injection surface where instructions hidden within bibliography entries could target the agent or its sub-agents during verification. Ingestion points: SKILL.md (Workflow Step 3 & 4), deep-verify.md (Sub-agent logic). Boundary markers: None identified; content is processed directly for key extraction. Capability inventory: Shell command execution (paperpile, scholarly), file system write access, and the ability to spawn parallel sub-agents. Sanitization: Not explicitly specified in the parsing logic.
[COMMAND_EXECUTION]: The skill uses shell commands to interact with specialized CLIs including paperpile, scholarly, and cli-council. These tools are used to query reference databases, verify DOIs, and orchestrate multi-model reviews.
[EXTERNAL_DOWNLOADS]: Retrieves academic metadata from well-known services such as Crossref (api.crossref.org) and OpenAlex. These network calls are scoped to retrieving citation data and do not involve the execution of remote code.

bib-validate