pipa-retrospective

Pass

Audited by Gen Agent Trust Hub on Aug 6, 2026

Risk Level: SAFE
Full Analysis
  • [DATA_EXPOSURE]: The skill is instructed to read configuration files located at ~/.pipa/communication-style.md and ~/.pipa/CONNECTORS.md. While these are application-specific paths, users should ensure these files do not contain sensitive credentials that could be inadvertently exposed if the agent includes their contents in its output.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from multiple external sources, creating an attack surface where malicious instructions in logs or feedback could influence agent behavior.
  • Ingestion points: Processes status history, closure artifacts, incident logs, retro notes, stakeholder feedback, and live records via composio-mcp.
  • Boundary markers: No explicit structural delimiters (such as XML tags or unique markers) are defined in the instructions to separate untrusted data from system prompts.
  • Capability inventory: The skill possesses the ability to perform external writes to playbooks or processes, but this capability is restricted by a requirement for explicit human approval.
  • Sanitization: No specific validation, escaping, or filtering of external content is defined in the instructions.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 6, 2026, 08:06 AM
Security Audit — agent-trust-hub — pipa-retrospective