fix-failed-paperclip-run

Pass

Audited by Gen Agent Trust Hub on Nov 16, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data—including failure reports, system logs, and execution traces—which it uses to diagnose errors and implement code repairs. This creates a vulnerability surface where malicious instructions hidden in logs or report metadata could influence the agent's behavior during the repair or PR creation process.\n
  • Ingestion points: The skill gathers evidence from supplied task URLs, failure reports, logs, and controller/provider traces as described in Section 1 and Section 2 of SKILL.md.\n
  • Boundary markers: There are no explicit instructions to wrap ingested data in delimiters or to ignore embedded instructions within the gathered evidence.\n
  • Capability inventory: The skill is authorized to read repository code, write code fixes, execute verification tests using the verify-paperclip tool, and deliver pull requests via prepare-paperclip-pr.\n
  • Sanitization: While the skill mandates the redaction of credentials and secrets from logs and notes, it does not specify methods for sanitizing or escaping the content of external reports before processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 16, 2026, 12:29 AM