Atlas / MCP servers / blueprint-conformance / bce-engine

bce-engineBLOCK

Name: bce-engine
Author: blueprint-conformance

mcp/blueprint-conformance/bce

MCP server by blueprint-conformance · agent skills · ai agents · apache 2 0 · Audited 2026-10-11

⚑ Report / re-scan Source ↗

Blueprints with Teeth — the blueprint conformance engine: authored architectural intent, enforced in CI.

Verdict

BLOCK

Grade

Trust score

21 /100

Exposed tools

13 10r · 3w · 0d

Transport

—

License

Apache-2.0

Stars

Overview

From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.

Current registry release: v0.5.0. Verify the exact public identity with npm view [email protected] version dist.integrity before installation. The same exact version is active as the Lane-A trust anchor.

The release retains the stack foundation (stack snapshot, stack diff, cross-OS golden-byte proof, and read-only MCP classification) and adds selected declared-stack enforcement to the ordinary gate, MCP, and offline v2 replay. It does not add automatic repair or registry resolution.

bce gives coding agents a local, deterministic architecture check. A versioned EngineeringBlueprint records the repository's structural rules. Agents draft those rules from your intent, inspect the code, and repair violations; the same engine checks every pull request. Humans own the intent and approve policy changes. Agents operate the day-to-day loop.

Released support (`v0.5.0`): TypeScript/JavaScript framework-surface AST extraction, direct TypeScript/JavaScript module boundaries, a Python import-surface MVP, and structured Python module boundaries. Node 22+ is required; the contract remains pre-1.0.

Evidence boundary: a score of 100 means the implemented checks passed over the observed surface. Route guard checks find governed call sites; they do not prove authorization, tenant isolation, or execution on every path. See the route evidence limits and correction, including the export forms missed by the immutable v0.3.0 release. Start with a bounded pilot and verify realistic

Read from source at commit b5d601c5b812OBSERVED · 2026-10-11

Connect

Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.

claude-code

claude mcp add workspace-access-lab --env BCE_GITHUB_TOKEN=${BCE_GITHUB_TOKEN} --env GH_TOKEN=${GH_TOKEN} --env GITHUB_TOKEN=${GITHUB_TOKEN} --env OPENAI_API_KEY=${OPENAI_API_KEY} -- npx -y workspace-access-lab

claude-desktop

{
  "mcpServers": {
    "workspace-access-lab": {
      "command": "npx",
      "args": [
        "-y",
        "workspace-access-lab"
      ],
      "env": {
        "BCE_GITHUB_TOKEN": "${BCE_GITHUB_TOKEN}",
        "GH_TOKEN": "${GH_TOKEN}",
        "GITHUB_TOKEN": "${GITHUB_TOKEN}",
        "OPENAI_API_KEY": "${OPENAI_API_KEY}"
      }
    }
  }
}

Exposed tools (13)

10 read · 3 write · 0 destructive.

Tool	Risk	Description
`assess_teeth`	read	Grade a blueprint’s TEETH (the toothlessness gate,
`check_baseline`	read	Read-only baseline maintenance check. Identifies removable debt and unaccepted new violations without changing policy.
`compare_blueprint_policy`	read	Read-only semantic comparison of a human-selected base and exact candidate blueprint; unknown direction blocks approval.
`doctor_repository`	read	Read-only adoption/readiness audit. Returns typed ready, needs-action, or refusal checks; mutates nothing.
`explain_constraint`	read	Read-only Promise/Lens/Proof/Limits explanation for one constraint using the canonical review function.
`get_report`	read	Read a serialized report JSON document the engine already wrote (a compliance-report.json, a
`greeting`	read	Fetch a greeting through the governed host.
`inspect_blueprint`	read	Read-only Promise/Lens/Proof/Limits inspection of an authored blueprint against a live repository tree.
`luna_chat`	write	Send a chat turn to the Luna assistant through the governed router.
`run_gate`	write	Call the real BCE MCP run_gate done-check for the current repository after making changes.
`stack_diff`	write	Read-only classification of every move between two committed StackManifests (added / removed /
`validate_blueprint`	read	Schema-validate an authored EngineeringBlueprint JSON file (the SAME parse the
`verify_review_packet`	read	Read-only replay of every deterministic packet binding and optional human decision binding; mutates nothing.

Trust audit

BLOCKgrade F · trust 21/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.

Layer	What it checks	Result
L0	Provenance & inventory	FAIL
L1	Static analysis of the code	FAIL
L2	Instruction surface (what it tells the agent)	FAIL
L3	Class-specific surface	PASS
L4	Behavioural (sandbox)	SKIPPED

What the source does

Filesystem: declared (11 observation(s))
Network: declared (5 observation(s))
Shell: declared (9 observation(s))
Dependencies: not all pinned
Secrets in source: found

Findings (25)

HIGHInventory / provenance · inv.suspicious_name · CWE-1104

assets/diagrams/c2-required-dependency-mobile.svg

c2-required-dependency-mobile.svg

Why it matters. member named after an attack tool

Fix. remove or justify

HIGHInventory / provenance · inv.suspicious_name · CWE-1104

assets/diagrams/c2-required-dependency.svg

c2-required-dependency.svg

Why it matters. member named after an attack tool

Fix. remove or justify

HIGHInventory / provenance · inv.suspicious_name · CWE-1104

scripts/verify-release-payload.mjs

verify-release-payload.mjs

Why it matters. member named after an attack tool

Fix. remove or justify

HIGHInventory / provenance · inv.suspicious_name · CWE-1104

scripts/verify-release-payload.selftest.mjs

verify-release-payload.selftest.mjs

Why it matters. member named after an attack tool

Fix. remove or justify

HIGHCode injection · code.eval_exec · CWE-78, CWE-94, CWE-95

research/model-evaluation/attested-implementations/by-git-commit/d19d583adafd5128cae890999950ec1f9ac92426/scripts/run-model-evaluation.mjs:434

.map((value) => `(allow process-exec (subpath ${sandboxLiteral(value)}))`);

Why it matters. evaluates text as code

Fix. remove; use a parser or a dispatch table

HIGHCode injection · code.eval_exec · CWE-78, CWE-94, CWE-95

research/model-evaluation/attested-implementations/by-git-commit/d19d583adafd5128cae890999950ec1f9ac92426/scripts/run-model-evaluation.mjs:478

.map((value) => `(allow process-exec (subpath ${sandboxLiteral(value)}))`);

Why it matters. evaluates text as code

Fix. remove; use a parser or a dispatch table

HIGHCode injection · code.eval_exec · CWE-78, CWE-94, CWE-95

scripts/run-model-evaluation.mjs:474

.map((value) => `(allow process-exec (subpath ${sandboxLiteral(value)}))`);

Why it matters. evaluates text as code

Fix. remove; use a parser or a dispatch table

HIGHCode injection · code.eval_exec · CWE-78, CWE-94, CWE-95

scripts/run-model-evaluation.mjs:518

.map((value) => `(allow process-exec (subpath ${sandboxLiteral(value)}))`);

Why it matters. evaluates text as code

Fix. remove; use a parser or a dispatch table

HIGHFilesystem / path · fs.credential_store · CWE-22, CWE-59

src/proposal-io.ts:20

/(^|\/)(?:id_rsa|id_ed25519|credentials|secrets?)(?:\.|$)/i,

Why it matters. touches a credential store

HIGHPrompt injection · prompt.override · CWE-94, CWE-1427

.agents/skills/skill-tuning/SKILL.md:125

| **S9** no-harness-imitation | Delete it. There is no legitimate reason for a distributed skill to carry text shaped like harness framing or an instruction to disregard prior instruction. |

Why it matters. asks the agent to drop prior instructions or safety