emws-api
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [DATA_EXFILTRATION]: The skill reads sensitive local files to facilitate API authentication and configuration.\n
- Evidence:
scripts/lib/config.tsaccesses~/.config/emws/.envfor secrets, andscripts/lib/auth.tsmanages cached tokens in~/.config/emws/tokens/.\n - Mitigation: A dedicated redactor (
scripts/lib/redact.ts) is used to mask these secrets in all console output and local log files. Additionally, the skill validates that the.envfile has restricted permissions (mode 600) before reading it.\n- [INDIRECT_PROMPT_INJECTION]: The skill processes data from external API responses, which could potentially contain adversarial instructions targeting the agent.\n - Ingestion points: API response bodies are retrieved in
scripts/lib/cli.tsand passed to the agent's context.\n - Boundary markers: The skill does not implement specific delimiters to isolate API content from the agent's instructions.\n
- Capability inventory: The agent can use the skill to perform HTTP requests (GET, POST, PUT, PATCH, DELETE) and record logs to the file system.\n
- Sanitization: While credentials are redacted, the textual content of API responses is not sanitized for natural language instructions.\n- [COMMAND_EXECUTION]: The skill utilizes local shell commands to gather metadata for its logging system.\n
- Evidence:
scripts/lib/log.tsexecutesgit branch --show-currentto record the current development context in the call logs.
Audit Metadata