dev-tooling
Pass
Audited by Gen Agent Trust Hub on Nov 9, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes a Python script (
scripts/detect.py) to execute various shell commands, includinggit,ddev, and version checks for installed binaries likeserena,golangci-lint, andphp. This is used for project reconnaissance to build the status table. - [EXTERNAL_DOWNLOADS]: The skill facilitates the installation of numerous development tools and libraries from standard package registries (NPM, PyPI, Composer) and the author's own skill repository (
hussainweb/skills). It also references the well-known Grounded Docs MCP server on GitHub. - [INDIRECT_PROMPT_INJECTION]: The skill possesses an attack surface for indirect prompt injection because it ingests data from numerous project configuration files (e.g.,
package.json,composer.json,pyproject.toml) and scrapes external web documentation. - Ingestion points:
scripts/detect.py(reads manifest and lock files),references/docs-index.md(scrapes external web documentation via MCP). - Boundary markers: Absent. The skill does not explicitly use delimiters to wrap ingested file content.
- Capability inventory: Uses the
Bashtool to run package managers and linter binaries, along withWriteandEdittools to modify project configurations. - Sanitization: Absent. Content is processed for tool status reporting and configuration generation.
- [DYNAMIC_EXECUTION]: The skill uses
uvxandnpxto dynamically fetch and execute tools such as Serena, Fallow, and Sentry Spotlight. This allows the skill to operate without requiring all tools to be pre-installed on the host system, which is standard practice for modern developer tools.
Audit Metadata