skillforge
PassAudited by ClawScan on May 1, 2026.
Overview
This instruction-only skill coherently helps generate Agent Skill packages, with no evidence of hidden code, credentials, exfiltration, or automatic execution.
This skill appears safe to install as an instruction-only generator, but review any skill packages it creates before installing them, especially generated scripts or instructions that could persistently change agent behavior.
Findings (2)
Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.
The skill may strongly steer how the agent approaches skill generation, but the provided text does not show deceptive or hostile goal changes.
This frames the pipeline with a role/context prompt. It is aligned with generating skills and does not contain an explicit instruction to override higher-priority instructions, but users should ensure it remains subordinate to platform and user instructions.
Set this as the system prompt (or prepend to Step 1) before starting the pipeline
Use this as task-specific guidance only; do not let generated or referenced prompts override system, developer, or user instructions.
Generated skills could influence future agent behavior if installed, especially if they include scripts or broad instructions.
The skill creates new agent-skill artifacts that may later be installed or reused. This is the stated purpose, but generated skills and scripts become part of the agent supply chain and should be reviewed.
Outputs a complete skill directory with SKILL.md, scripts/, references/, and templates/.
Review generated SKILL.md files and any generated scripts before installing, sharing, or running them.