agency-reality-checker
Pass
Audited by Gen Agent Trust Hub on Mar 25, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes several shell commands to inspect the local filesystem and project source code. Evidence includes the use of
ls -lato verify directory structures andgrepto search for specific UI implementation keywords across HTML, CSS, and Blade templates. - [UNVERIFIABLE_DEPENDENCIES]: The agent is instructed to execute a local shell script named
./qa-playwright-capture.sh. Since this script is not provided in the skill's file list, its behavior is unverifiable and could execute arbitrary code if the script present in the environment is malicious. - [INDIRECT_PROMPT_INJECTION]: The skill's workflow involves reading and analyzing automated test results, which creates a potential surface for indirect instructions to influence agent behavior.
- Ingestion points: Reading of
public/qa-screenshots/test-results.jsonvia thecatcommand in SKILL.md. - Boundary markers: Absent. No explicit instructions are given to ignore embedded prompts within the test data.
- Capability inventory: The skill has access to shell command execution (
ls,grep,cat, and the local playwright script). - Sanitization: Absent. The data from the test results is used directly to generate the final integration report without validation.
Audit Metadata