session-debug

Pass

Audited by Gen Agent Trust Hub on Aug 29, 2026

Risk Level: SAFECREDENTIALS_UNSAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [Sensitive File Access]: The instructions guide the agent to check ~/.amplifier/keys.env. This file path is commonly used to store sensitive API keys and environment variables, and accessing it presents a potential risk of credential exposure.
  • [Indirect Prompt Injection Surface]: The skill is designed to process external data from session logs (events.jsonl) and user arguments, which could contain instructions that influence the agent's diagnostic behavior.
  • Ingestion points: The skill reads from the events.jsonl log file and the $ARGUMENTS variable in SKILL.md.
  • Boundary markers: No specific delimiters or "ignore instructions" warnings are defined for the ingested log content.
  • Capability inventory: The skill utilizes a delegation tool to invoke a specialized session-analyst agent.
  • Sanitization: There are no explicit steps described for sanitizing or filtering the content of the logs before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 29, 2026, 05:21 AM
Security Audit — agent-trust-hub — session-debug