agents-md-generator
Pass
Audited by Gen Agent Trust Hub on Oct 2, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONPRIVILEGE_ESCALATION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes a bundled detector script (
detect-agent-context) to scan the repository for manifests and configuration files. This script is provided in multiple formats (Python, Shell, PowerShell) to ensure portability. - [PRIVILEGE_ESCALATION]: The Windows batch wrapper (
detect-agent-context.cmd) utilizes the PowerShell-ExecutionPolicy Bypassflag to run its internal script. This is a common practice to ensure functionality on systems with restricted default policies and is scoped specifically to the skill's execution. - [INDIRECT_PROMPT_INJECTION]: The skill processes project manifests and existing instruction files to generate new instructions, creating an attack surface for indirect prompt injection from repository content.
- Ingestion points: The skill reads project manifests (
package.json,pyproject.toml, etc.) and existing instruction files (AGENTS.md,CLAUDE.md) using thedetect-agent-contextscripts and instructions inSKILL.md. - Boundary markers: The instructions do not define specific delimiters for processed untrusted content but provide logic for selecting only "durable" and "verified" facts.
- Capability inventory: The skill has the capability to execute local scripts and write or modify Markdown instruction files within the repository.
- Sanitization: The skill includes explicit compaction rules to remove role-play personas, generic advice, and stale guidance from the generated output.
Audit Metadata