takt-optimize
Pass
Audited by Gen Agent Trust Hub on Mar 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes a shell script
scripts/validate-takt-files.shto verify optimized files. This script contains a relative path reference../../takt-piece/scripts/validate-takt-files.sh, which relies on a specific external directory structure being present in the agent's environment. - [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface due to its data ingestion and file generation capabilities.
- Ingestion points: In Step 1 and Step 7, the skill reads untrusted data from
.takt/logs/*.jsonl(execution logs) and user-provided piece/facet files located in~/.takt/pieces/or.takt/pieces/. - Boundary markers: No explicit delimiters or instructions to ignore embedded directives within the ingested logs or facets are specified in the prompt logic.
- Capability inventory: The skill possesses the ability to create, delete, and modify files (Step 3) and execute shell scripts (Validation step).
- Sanitization: The skill lacks a sanitization layer for the ingested data; instructions from the logs or existing facets could potentially influence the logic of the newly generated files or the parameters of the validation command.
Audit Metadata