The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands (specifically grep) and inline Python scripts as 'mechanical gates' during the final phase of paper preparation. These commands are used to verify the absence of placeholders (e.g., [TBD], [PENDING]) and 'AI-tell' phrases in the generated paper_draft.md.
Evidence: Found in SKILL.md (Phase 5) and references/QUALITY_CHECKLIST.md.
Note: These scripts operate locally on the skill's own output and are intended for quality assurance.
[EXTERNAL_DOWNLOADS]: The skill references the Crossref REST API (api.crossref.org) to verify the existence and metadata of scholarly citations.
Evidence: Found in references/CITATION_INTEGRITY.md ('WebFetch https://api.crossref.org/works/').
Note: Crossref is a standard, well-known service for academic metadata; this behavior is consistent with the skill's goal of citation integrity.
[COMMAND_EXECUTION]: A Python test script (tests/test_document_contract.py) uses subprocess.run to execute the system grep binary.
Evidence: Found in tests/test_document_contract.py.
Note: This script is a unit test designed to ensure the regex patterns defined in the skill's instructions correctly identify restricted phrases when run by the agent's environment, preventing 'silent failures' of the verification gates.

research-paper