callidescope-trace

Pass

Audited by Gen Agent Trust Hub on Sep 8, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructs the agent to execute npx callidescope, which downloads and runs an external package from the NPM registry.
  • [INDIRECT_PROMPT_INJECTION]: The skill has a vulnerability surface for indirect prompt injection because it ingests untrusted data from the workspace code and presents it to the agent.
  • Ingestion points: The tool reads TypeScript source files, specifically signatures and JSDoc comments, as described in SKILL.md.
  • Boundary markers: There are no explicit instructions or delimiters defined to warn the agent to ignore instructions embedded within the analyzed code snippets.
  • Capability inventory: The skill uses npx to execute commands and reads various workspace files (TS, JSON, Markdown).
  • Sanitization: While the skill mentions collapsing JSDoc prose to a single line and trimming long signatures, there is no evidence of sanitization or filtering to prevent malicious instructions within those comments from influencing the agent's behavior.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 8, 2026, 01:11 PM
Security Audit — agent-trust-hub — callidescope-trace