Least-privilege governance for ERPNext — the credential floor + a governed agent broker, MCP · A2A · API, one spine. No debit without a credit: PLAN · CONSENT · PROVE · UNDO.
- vulnerabilities attested + 32
- capability exposure inferred + 14
- recent drift inferred + 12
- tool safety inferred + 5
- trust mitigators mixed − 3
attested inferred mixed
The A–E grade is our heuristic synthesis — a "review this" prompt, not a verdict. Each factor is tagged by what backs it: attested (a verifiable record), reported (a third party's claim), or inferred (our own heuristic, e.g. permissions). See methodology.
graded 3m ago · see ecosystem CVEs →
- C · 43 → D · 60
- A · 11 → C · 43
- A · 0 → A · 11
- medium dangerous code
dynamic exec: eval()/exec()
analyzed commit 10b45ce · analyzer v28 · 1d ago
danger signals1
- dynamic code execution eval()/exec() john-broadway-pacioli-10b45ce/guard/pacioli_guard/tests/test_act.py :383
exec(compile( # noqa: S102 — building a frame double, not evaluating input
- recent drift +12 capability drift →
[](https://mcpobservatory.com/servers/github:john-broadway/pacioli/security) Heuristic, inferred signals — false positives (legitimately powerful tools, forks, language ports) are expected. Treat each as "review this", not a verdict. See the ecosystem-wide picture on the security hub, or the fleet security of john-broadway.