strategy-interview

Pass

Audited by Gen Agent Trust Hub on Sep 20, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [NO_CODE]: The skill consists entirely of Markdown instruction and reference files. It contains no executable scripts (e.g., Python, JavaScript, or Shell) that would run on the host system.
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze user-provided strategy documents and existing interview state. This represents a standard surface for indirect prompt injection, which is managed here by the skill's core purpose of critical analysis.
  • Ingestion points: User-provided text, existing strategy-notes.md files, and persisted state files within the .beagle/strategy/ directory.
  • Boundary markers: The skill does not explicitly use delimiters when reading user documents, relying instead on the agent's context to separate instructions from analyzed content.
  • Capability inventory: The skill can write files to the working directory and the .beagle/ sub-directory for state persistence.
  • Sanitization: The skill implements a semantic "bad-strategy filter" to detect logical fallacies and buzzwords, which serves as a natural defensive layer against low-quality or manipulative input.
  • [PERSISTENCE]: The skill uses a local .beagle/strategy/ directory to maintain "Durable interview state." This is a legitimate state-management feature for long-running conversations and does not involve malicious persistence mechanisms like modifying system startup scripts.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 20, 2026, 09:16 AM
Security Audit — agent-trust-hub — strategy-interview