inno-idea-generation

Pass

Audited by Gen Agent Trust Hub on Sep 21, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process untrusted data from external sources and project files, which introduces a potential vulnerability to indirect prompt injection.
  • Ingestion points: The skill uses WebSearch, WebFetch, read_file, and search_project to gather information from the internet and the local workspace (specified in SKILL.md).
  • Boundary markers: There are no instructions to use delimiters or explicit warnings for the agent to ignore instructions contained within the ingested external data.
  • Capability inventory: The skill has access to sensitive capabilities including write_file and run_terminal (specified in SKILL.md), which could be misused if malicious instructions are successfully injected via processed content.
  • Sanitization: No evidence of input validation or sanitization of the fetched external content was found in the instructions.
  • [COMMAND_EXECUTION]: The skill executes bundled shell scripts (scripts/scamper-prompts.sh, scripts/swot-template.sh, scripts/research-sources.sh) using the run_terminal tool to generate templates and brainstorming prompts. The scripts themselves perform safe text output, but the capability to execute bundled scripts should be monitored for changes in intent.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 21, 2026, 05:58 PM