update-trace-engine
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONDATA_EXFILTRATIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes several local shell scripts and JavaScript files to build, test, and validate the trace engine package. Examples include
prep-trace-engine-package.sh,copy-build-trace-engine-for-publish.sh, and Lighthouse-specific test suites usingmochaandtsc. - [DATA_EXFILTRATION]: The skill performs network operations via Git and npm. It fetches from and pushes to repositories on
chromium.googlesource.comandgithub.com, and interacts with the officialregistry.npmjs.org. These operations are transparent and consistent with the skill's stated purpose of rolling a package dependency. - [INDIRECT_PROMPT_INJECTION]: The skill represents an attack surface by ingesting data from external repositories and registries to generate PR descriptions and automate versioning.
- Ingestion points: External data enters the agent context via the output of
git log(upstream commit history) andnpm view(registry version information). - Boundary markers: No explicit boundary markers or "ignore" instructions are used when interpolating this data into shell command arguments or PR templates.
- Capability inventory: The skill has the capability to execute shell commands, write to the local file system (modifying
.d.tsfiles), and perform network operations via Git/npm. - Sanitization: There is no explicit sanitization of the data retrieved from
git logornpm viewbefore it is used in subsequent command arguments.
Audit Metadata