The Agent Skills Directory

[DATA_EXFILTRATION]: The scripts/api_request.py tool performs network operations to external marketing systems. It mitigates exfiltration risks by automatically redacting sensitive strings (e.g., 'api-key', 'secret', 'token', 'authorization') from URLs, headers, and logs. It strictly enforces HTTPS for remote connections and requires credentials to be supplied via environment variables rather than command-line arguments to prevent leakage in process lists or logs.
[COMMAND_EXECUTION]: The scripts/test_api_request.py script executes the local api_request.py file using subprocess.run to perform automated unit testing. This execution is restricted to the local environment, avoids shell interpolation, and targets a verified, internal script path.
[INDIRECT_PROMPT_INJECTION]: The skill handles data from external API responses, which could potentially contain malicious instructions. Ingestion points: API response bodies processed in scripts/api_request.py. Boundary markers: Defined in references/adapter-contract.md, including dry-run and external state verification protocols. Capability inventory: Ability to perform HTTP mutations (POST, PUT, DELETE) via api_request.py when explicitly authorized. Sanitization: Implementation of a 5MB response size limit and strict JSON object parsing to limit the impact of large or malformed payloads.

tool-connectors