ksp
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONMETADATA_POISONINGINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a bundled Python script (
scripts/inspect-project.py) to analyze the target repository's configuration. The script performs local file reads and regex-based analysis to generate a KSP compatibility report. - [METADATA_POISONING]: The skill's YAML frontmatter contains deceptive metadata, including future timestamps (2026) and references to non-existent or hallucinated AI model versions (
gpt-5.6-sol,gpt-6.1-sol). While not directly harmful, this information is inaccurate. - [INDIRECT_PROMPT_INJECTION]: The skill possesses an indirect prompt injection attack surface.
- Ingestion points: The
inspect-project.pyscript recursively reads and analyzes local project files includingbuild.gradle,.gradle.kts,settings.gradle,gradle.properties,libs.versions.toml,.kt, and.javafiles. - Boundary markers: The script does not utilize explicit boundary markers or security warnings when outputting the content or analysis of these files to the agent.
- Capability inventory: The skill allows the agent to read project files, run an inspection script, and provide instructions for code generation.
- Sanitization: There is no evidence of sanitization or escaping of the ingested file content before it is processed by the agent in Step 2.
Audit Metadata