levelup-clarify
Pass
Audited by Gen Agent Trust Hub on Aug 17, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local setup scripts (
scripts/bash/setup-levelup-clarify.shorscripts/powershell/setup-levelup-clarify.ps1) to resolve the project root and count pending CDR files using system utilities like git and grep. - [PROMPT_INJECTION]: The skill processes untrusted content from Context Directive Records, representing a surface for indirect prompt injection.
- Ingestion points: Reads CDR files from the project's
.adlc/drafts/cdr/directory and interpolates their context, decision, and evidence fields into the prompt. - Boundary markers: While using markdown headers, the skill lacks explicit delimiters or instructions to ignore embedded commands within the CDR fields, which could allow malicious content in a CDR to influence agent behavior.
- Capability inventory: The agent has the ability to execute scripts, write to the filesystem (updating records and indices), and initiate LLM-based regression tests.
- Sanitization: No sanitization or filtering is applied to the content of the CDR files before they are presented to the agent or interpolated into instructions.
Audit Metadata