The Agent Skills Directory

[SAFE]: The skill utilizes vendor-specific infrastructure for legitimate configuration and performance tracking.
The script scripts/lib/hosted-api.ts communicates with api.forsvn.com to fetch configuration packs and post cycle metrics.
It manages its own API keys stored in ~/.config/forsvn/credentials.json, which is a standard pattern for the toolchain associated with the author hungv47.
[PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by processing untrusted data from external sources.
The Metric Ingest Agent processes raw reply threads and CRM data which could contain adversarial instructions.
The skill possesses filesystem write capabilities to update ledgers and evaluation artifacts.
Evidence of mitigation: Support scripts such as append-loop-result.ts include validation to prevent tab or newline injections in data fields, and file operations are protected by symbolic link checks and path sanitization.

evaluate-outreach