axiom-optimize-build
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from Xcode project files to determine build settings and optimization opportunities.
- Ingestion points: Reads project configurations, targets, and build phases using
xcproject inspectand searches for patterns inPackage.swiftand.xcodeprojfiles. - Boundary markers: There are no explicit delimiters or instructions to ignore embedded commands within the project files being analyzed.
- Capability inventory: The agent has the capability to run
xcodebuildcommands and use system utilities likesysctl,pgrep, andgrep. - Sanitization: While the skill uses shell quoting for project-related variables, it does not perform deep validation of the contents of the project configuration files it processes.
- [COMMAND_EXECUTION]: The skill uses system-level commands to perform build analysis and environment checks.
- Evidence: Executes
xcodebuildwith custom compiler flags (-Xfrontend -debug-time-function-bodies) to generate timing data for functions and expressions. - Evidence: Uses
sysctl -n hw.ncputo determine available hardware resources andpgrepto monitor active build processes.
Audit Metadata