reagent-expiry-alert
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-supplied strings for reagent barcodes, names, and locations, creating a surface for potential instruction injection when these values are included in generated reports.\n
- Ingestion points: User input via
--scan,--name, and--locationparameters as documented inSKILL.md.\n - Boundary markers: Absent; user-provided data is interpolated directly into report templates without delimiting instructions.\n
- Capability inventory: The
scripts/main.pyscript performs file-write operations to a local JSON inventory file (~/.openclaw/reagent_inventory.json).\n - Sanitization: Absent; string inputs are not validated or escaped before storage or display.\n- [INDIRECT_PROMPT_INJECTION]: The skill bundle includes a self-referential audit report (
reagent-expiry-alert_audit_result_v2.json) containing safety and security verdicts, which should not be considered authoritative.\n - Ingestion points: The
reagent-expiry-alert_audit_result_v2.jsonfile is part of the skill's file set.\n - Boundary markers: Not applicable.\n
- Capability inventory: Not applicable.\n
- Sanitization: Not applicable.\n- [EXTERNAL_DOWNLOADS]: The skill references the official source repository on GitHub for its code base.\n
- Source:
https://github.com/aipoch/medical-research-skills\n - Context: Documented as the source location in
SKILL.md.
Audit Metadata