jestatsio/toolprint scan

Scan MCP servers for tool poisoning, secret leaks, and silent rug-pulls; enforce the committed toolprint.lock.

Trust Signals

name	description	required	default
target	MCP config path, http(s) URL, npx:<package>, or command. Optional if a config is auto-discovered or --config is set.	no	""
config	Path to an MCP client config file (Claude Desktop / VS Code / Cursor).	no	""
fail-on	Minimum severity that fails the build: info\|low\|medium\|high\|critical.	no	high
lockfile	Path to the toolprint.lock to enforce.	no	""
sarif-file	If set, write SARIF 2.1.0 results to this path for github/codeql-action/upload-sarif. In SARIF mode findings do not fail the job (they become code-scanning alerts); operational errors still do. Upload with `if: always()`.	no	""
version	toolprint version to run (npm dist-tag or exact version).	no	latest
comment-on-pr	On a pull_request event (non-SARIF mode), upsert a sticky comment summarizing the scan. Requires `pull-requests: write` permission. Has no effect when sarif-file is set.	no	false
github-token	Token used to post the PR comment. Defaults to the workflow token.	no	${{ github.token }}

no outputs