skill-suggester
Pass
Audited by Gen Agent Trust Hub on Oct 27, 2026
Risk Level: SAFEDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [DATA_EXFILTRATION]: The skill accesses interaction history files located at ~/.local/state/opencode/prompt-history*.jsonl. These files represent a significant data exposure risk as they commonly contain sensitive information, including API keys, passwords, private identifiers, and intellectual property discussed in previous sessions.- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection by processing untrusted historical data.
- Ingestion points: Historical prompt data is read from local JSONL files.
- Boundary markers: The skill instructions do not define boundary markers or instructions to isolate the historical content from the current execution context.
- Capability inventory: The skill utilizes file read access and prompt interpolation to analyze and display past content.
- Sanitization: There is no evidence of sanitization or filtering applied to the historical messages before they are processed by the agent. A malicious instruction embedded in a past prompt could be re-activated when the skill re-processes that content.
Audit Metadata