skills/datadrivenconstruction/ddc_skills_for_ai_agents_in_construction/json-parser/Gen Agent Trust Hub
json-parser
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from external sources (JSON strings and files) which can influence agent behavior if the data contains embedded instructions.
- Ingestion points: The ConstructionJSONParser class in SKILL.md includes parse_file(file_path) and parse_string(json_string) methods that read external data into the agent's context.
- Boundary markers: The instructions do not define specific delimiters or 'ignore instructions' wrappers for the ingested JSON content.
- Capability inventory: The skill has the capability to read local files via the open() function in SKILL.md, which is a permission requested in claw.json. It does not perform network operations or execute code based on the data.
- Sanitization: The skill uses standard json.load() and json.loads() for parsing, which prevents basic code injection during the parsing phase but does not filter natural language instructions embedded within the JSON values.
Audit Metadata