The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes data from external JSON report files, which creates an attack surface for indirect prompt injection if the inputs contain malicious instructions targeting a downstream agent. 1. Ingestion points: The skill accepts three external JSON report files through CLI arguments --detector-json, --news-json, and --price-action-json in scripts/run_contrarian_setup_gate.py. 2. Boundary markers: No explicit delimiter boundaries or 'ignore' instructions are provided to the agent for the output reports. 3. Capability inventory: The skill writes JSON and Markdown reports to the local file system. 4. Sanitization: The skill implements strict schema validation, type checking, and a specialized scanner in gate_logic.py that rejects any input containing non-finite floats to prevent malformed output generation.
[DYNAMIC_EXECUTION]: The skill's test infrastructure uses dynamic module loading to verify consistency across different components of the workflow. Evidence: The script scripts/tests/test_gate_logic.py utilizes importlib.util to dynamically load and execute the weekly_price_action.py script from a relative path outside the skill's own directory.

contrarian-setup-gate