setup-pstack

Pass

Audited by Gen Agent Trust Hub on Sep 2, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests data from a local configuration file located at ~/.agents/pstack-models.md. While this presents a surface for indirect prompt injection if the file is modified by an attacker, the skill mitigates this by validating model slugs against a list of confirmed available models before writing the config.
  • Ingestion points: Reads ~/.agents/pstack-models.md in Step 2.
  • Boundary markers: Not explicitly defined in the file format, but the skill uses a specific line-based role mapping.
  • Capability inventory: File system write access to the specific configuration path and invocation of the /create-verification-skill command.
  • Sanitization: Implements model slug validation in Step 4 to ensure only legitimate model identifiers are processed.
  • [DYNAMIC_EXECUTION]: Step 7 describes the ability to invoke /create-verification-skill to generate project-local scripts. This represents a managed form of code generation typical for setup and development tools, where the logic is delegated to a dedicated skill rather than executing arbitrary strings directly.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 2, 2026, 06:53 PM
Security Audit — agent-trust-hub — setup-pstack