moai-change-logger
Pass
Audited by Gen Agent Trust Hub on Mar 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill requests the
Bashtool and describes logic for trackingcommands_executedvia functions liketrack_command_history(). This enables the agent to monitor and record shell activity for auditing purposes. - [DATA_EXFILTRATION]: The skill utilizes
Read,Bash, andWebFetchtools to aggregate sensitive local data including file contents, Git history, and developer activity logs. While this is used for change analytics, the capability to read local logs and access the network creates a surface for potential data exposure. - [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by ingesting and processing untrusted data from the local environment.
- Ingestion points: Monitored file contents, Git commit messages, and shell command history logs.
- Boundary markers: There are no documented delimiters or instructions used to separate monitored content from agent instructions in the provided snippets.
- Capability inventory: The skill has access to
Bash,Write,WebFetch, andReadtools. - Sanitization: No validation or escaping of the ingested log data is demonstrated in the implementation patterns.
Audit Metadata