The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external data sources such as CSV, Excel, and JSON files, which introduces a potential surface for indirect prompt injection if those files contain malicious instructions.
Ingestion points: Data enters the system via ingest_batch and ingest_reading methods, as well as through user-provided files mentioned in instructions.md.
Boundary markers: The provided instructions do not specify explicit delimiters or "ignore instructions" wrappers for the ingested data.
Capability inventory: The skill uses internal logic for statistical analysis and reporting. The claw.json file requests filesystem permissions, allowing the agent to read and write files.
Sanitization: The Python implementation performs numerical validation (physical limit checks and Z-score anomaly detection), but lacks specific text sanitization if string metadata from readings were to be interpolated into agent prompts.
[SAFE]: The external domain datadrivenconstruction.io is the official homepage for the author datadrivenconstruction and is used legitimately for documentation and project identification.

sensor-data-aggregator