flux-opencode-improve-agents-md
Pass
Audited by Gen Agent Trust Hub on Jul 31, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill invokes a local project utility named
fluxctllocated in the.opencode/bin/directory. This is used for internal session tracking and does not involve external network calls or privileged operations. - [PROMPT_INJECTION]: The skill processes
AGENTS.mdfiles, which introduces an indirect prompt injection surface. Malicious instructions inside the processed documentation could theoretically attempt to influence the agent's behavior during the restructuring task. - Ingestion points: Loads
AGENTS.mdor a user-specified path from the skill arguments. - Boundary markers: Explicitly preserves
<!-- BEGIN FLUX -->markers and foundational sections. - Capability inventory: Includes file system read/write access to project documentation and execution of local workspace binaries.
- Sanitization: Restructuring is guided by specific rules and target tags, but the ingested content is not sanitized against embedded instructions.
Audit Metadata