The Agent Skills Directory

[EXTERNAL_DOWNLOADS] (MEDIUM): The installation instructions (SKILL.md) recommend downloading and executing code from an untrusted GitHub repository (github.com/barneyjm/camino-skills) using npx skills add and clawhub. This source is not on the trusted organization or repository list.
[DATA_EXFILTRATION] (LOW): The script scripts/context.sh transmits user-provided location data and an API key (CAMINO_API_KEY) to an external domain (api.getcamino.ai). This domain is not on the pre-approved whitelist for exfiltration analysis.
[COMMAND_EXECUTION] (SAFE): The shell script uses curl and jq with properly quoted variables ("$INPUT"), which prevents common shell injection vulnerabilities when handling the JSON payload.
[INDIRECT_PROMPT_INJECTION] (LOW): The skill has a data ingestion surface via command-line arguments.
Ingestion points: The scripts/context.sh script accepts arbitrary JSON input.
Boundary markers: None present in the script or instructions.
Capability inventory: Limited to network requests via curl and JSON processing via jq.
Sanitization: The script performs basic JSON validation using jq empty before processing.

context