idor

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 1.00). The skill instructs the agent to include session cookies/tokens (e.g., "Cookie: session=YOUR_SESSION") and other credentials directly in generated curl commands, scripts, and return summaries, which requires handling and outputting secret values verbatim and therefore poses a high exfiltration risk.

CRITICAL E006: Malicious code pattern detected in skill scripts.

• Malicious code pattern detected (high risk: 1.00). This document is an explicit, actionable offensive pentest guide for exploiting IDOR/broken access control—containing step-by-step techniques for mass enumeration, automated data extraction, credential/session theft, privilege escalation, and file/backup retrieval—so it enables deliberate data exfiltration and abuse if used without authorization.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.65). The skill’s runtime workflow instructs the tester to fetch and compare responses from target endpoints using attacker-controlled IDs (e.g., curl ... /api/users/OTHER_ID/profile and GraphQL queries), and those responses (which may include outsider-authored user data) are ingested into the agent’s LLM context via the tool/output/evidence text the skill saves/prints.