The Agent Skills Directory

[SAFE]: The skill provides legitimate observability patterns for monitoring API usage. It uses standard TypeScript code to wrap Anthropic SDK calls and log metadata including token counts, latency, and estimated costs.
[INDIRECT_PROMPT_INJECTION]: The skill presents a standard ingestion surface for code modification tasks.
Ingestion points: The trackedCreate function processes Anthropic.MessageCreateParams and API responses in SKILL.md.
Boundary markers: None present in the code templates.
Capability inventory: The skill uses Read, Write, and Edit tools to integrate code into the user's project.
Sanitization: Standard for observability code, no specific sanitization is implemented for the metadata logs.
[DATA_EXPOSURE_&_EXFILTRATION]: Logging is performed exclusively via console.log and console.error. No sensitive file access or network exfiltration patterns were identified.

clade-observability