dstack-beads-project-alignment-execute

Pass

Audited by Gen Agent Trust Hub on Aug 29, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill invokes a local binary dstack using a relative path "{baseDir}/../../bin/dstack". This tool is used to manage and commit project 'alignment' corrections.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection by ingesting and acting upon untrusted external data.
  • Ingestion points: Data enters the context through the alignment claim-next command, which retrieves 'Beads' containing correction instructions.
  • Boundary markers: There are no explicit delimiters or instructions to the agent to treat the content of the Beads as untrusted data rather than instructions.
  • Capability inventory: The skill allows the agent to modify the file system (implementing corrections) and perform repository operations (git commit).
  • Sanitization: The skill includes a revalidation step to check if the finding is obsolete, but it does not specifically sanitize the input for malicious instructions targeting the LLM.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 29, 2026, 12:02 AM
Security Audit — agent-trust-hub — dstack-beads-project-alignment-execute