deslop
Pass
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local shell commands to manage repository state and verify code modifications.
- Evidence: It invokes
git diffandgit statusto define the editing scope andruns the narrowest covering testfor changed files or packages to ensure functional correctness in Step 4. - [PROMPT_INJECTION]: The skill ingests untrusted data from git diffs and source code, creating a surface for indirect prompt injection attacks.
- Ingestion points: Git diff outputs and source file contents are read during the inspection and classification phase (Step 2).
- Boundary markers: The skill does not explicitly specify delimiters or instructions to ignore embedded prompts within the processed code text.
- Capability inventory: The agent has the ability to modify the local filesystem (Step 3) and execute shell commands for testing (Step 4).
- Sanitization: No sanitization or validation of the input diff text is described in the instructions.
Audit Metadata