Reports & advisories
The pieces you reach for in a write-up: a shareable HTML advisory, and filled CVSS v3.1 + CWE markdown advisories.
# Render a recorded session (+ findings) into ONE self-contained HTML advisorymcpwn report session.jsonl --findings ./out -o report.html
# Turn findings into filled CVSS v3.1 + CWE advisories (GHSA-style markdown)mcpwn advisory ./out -o ./advisoriesmcpwn report
Section titled “mcpwn report”Turns a session log (the JSONL the proxy writes via --log) plus an optional
findings dir into a single zero-dependency HTML advisory: header stats,
findings by severity, side-by-side tamper diffs, and a tiny vanilla-JS severity
filter. It is a static file - nothing listens, fetches, or executes. It
opens offline and screenshots cleanly. Every byte of session/finding content is
HTML-escaped (the content is literally the attacker-controlled traffic under
test).
mcpwn advisory
Section titled “mcpwn advisory”Turns a findings dir / findings.json into filled CVSS v3.1 + CWE advisories
(GHSA-style markdown mirroring the
advisory template) - one .md per finding,
a combined doc, or stdout:
mcpwn advisory ./out -o ./advisories # one .md per findingmcpwn advisory ./out -o advisories.md # one combined documentmcpwn advisory ./out # combined document to stdoutCWE/CVSS also enrich the SARIF rule metadata and the HTML report. The CWE + CVSS
v3.1 scoring for each finding id lives in
src/advisory.ts.
See the threat taxonomy for the per-class mapping, and
disclose responsibly per the security policy.