x-undocumented-api
Pass
Audited by Gen Agent Trust Hub on Jul 21, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The script
scripts/extract-x-graphql-endpoints.tsfetches content fromx.comandabs.twimg.comto identify GraphQL operations and query IDs. These downloads target official service domains and are documented as part of the skill's discovery process. - [DATA_EXFILTRATION]: The skill facilitates the use of sensitive browser authentication materials, including
auth_tokenandct0cookies, and bearer tokens. It contains multiple explicit directives instructing the agent to redact these secrets and prevent them from appearing in any output or logs. - [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface because it processes external JavaScript bundles from X.com to extract API metadata.
- Ingestion points: Metadata is ingested via the
scripts/extract-x-graphql-endpoints.tsscript which fetches and parses third-party JavaScript files. - Boundary markers: The skill relies on behavioral instructions for the agent rather than structural delimiters (like XML tags) to isolate the ingested data.
- Capability inventory: The skill has the capability to perform network operations (live API probes) and file system writes (saving extracted metadata) across its scripts and instructions.
- Sanitization: The instructions recommend extracting only specific metadata and using secret-safe redacted formats for reporting findings.
Audit Metadata