Atlas / MCP servers / rigour-labs / Rigour

RigourBLOCK

Name: Rigour
Author: rigour-labs

mcp/rigour-labs/rigour

MCP server by rigour-labs · agentic workflow · ai · ai agents · Audited 2026-10-09

⚑ Report / re-scan Source ↗

The immune system for AI coding agents

Verdict

BLOCK

Grade

Trust score

50 /100

Exposed tools

45 33r · 11w · 1d

Transport

stdio

License

MIT

Stars

Overview

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

[](https://www.npmjs.com/package/@rigour-labs/cli) [](https://www.npmjs.com/package/@rigour-labs/cli) [](https://opensource.org/licenses/MIT) [](https://rigour.run)

A reviewer that knows how your team works, whichever agent writes the code.

Whichever agent your engineers use, its work arrives already shaped by how your team builds software, and it gets closer to your standards every week.

Teams used to carry their judgment through people: the senior who briefed you before you started, the reviewer who caught what you missed, the engineer who remembered why the last attempt failed. Coding agents write faster than any person can brief, review or remember. The knowledge is still in your team; Rigour is how it reaches every agent, at the speed agents work.

Works with Claude Code, Cursor, Codex, Cline and Windsurf. Free, open source, and runs on your machine.

How it does it

Brief before writing. Before an agent edits a file, it is told what your team asks of that file: the rules your repository wrote for it, the lessons your team learned on it, the points your team settled against. At most three items per file, each cited. Opt-in: rigour hooks init --brief, or rigour_brief for any agent with MCP. See The briefing.
Verify while writing and before push. Your team's checks and rules gate any agent's work: on every edit, before the agent says "done", and before git push, where your formatter, linter, type check and the tests that touch the change run too.
Review with your team's context, at your team's severity. A reviewer that reads every point from your last human review, checks each rule your repository

Read from source at commit 67ad92620e21OBSERVED · 2026-10-09

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 (npm)

claude mcp add mcp -- npx -y @rigour-labs/[email protected]

Exposed tools (45)

33 read · 11 write · 1 destructive. Blast radius: 1 tool can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.

Tool	Risk	Description
`cwd`	read	Absolute path to the project root.
`echo`	read	Echo input
`git`	write	Run a read-only git command in the repository: log, show, diff, blame, grep, ls-files, rev-parse, merge-base.
`grep`	read	Search tracked files with an extended regular expression. Returns path:line:text matches.
`hidden`	read	Not allowed
`list_dir`	read	List a directory in the repository or the review inputs.
`read_file`	read	Read lines of a repository file, numbered. Use it to check callers, callees, types and constants before claiming a defect.
`rigour-ai-health-report`	read	AI code health report focusing on drift detection: hallucinated imports, duplication drift, context window artifacts, inconsistent error handling, and promise safety.
`rigour-deep-analysis`	write	Run deep LLM-powered code quality analysis. AST extracts facts, LLM interprets patterns (SOLID violations, code smells, architecture issues), AST verifies findings. Local sidecar by default; cloud provider mode when configured.
`rigour-fix-loop`	read	Iteratively fix all quality gate violations until the project passes. Retrieves fix packets and resolves issues in priority order (critical → low).
`rigour-pre-commit`	write	Pre-commit review: Rigour
`rigour-security-review`	read	Full security review: CVE audit on dependencies + code-level vulnerability scan (OWASP LLM Top 10). Reports all findings with remediation steps.
`rigour-setup`	read	Initialize Rigour quality gates for a project. Runs gate checks, installs IDE hooks, and reports the initial quality score breakdown.
`rigour_agent_deregister`	read	Deregister an agent from the multi-agent session. Use when an agent completes its work or needs to release its scope for another agent.
`rigour_agent_register`	write	Register an agent in a multi-agent session. Use this at the START of agent execution to claim task scope and enable cross-agent conflict detection. Required for Agent Team Governance.
`rigour_brief`	read	Before writing: with goal, the team
`rigour_cache_stats`	read	Returns detailed performance stats across all 4 cache layers (exact hits, semantic hits, partial hits, misses, hit rate, tokens served from cache).
`rigour_check`	write	Run quality gate checks on the project. MUST be called before declaring any coding task complete. Checks code complexity, file size, required docs, security patterns, and more. Returns PASS or FAIL with details. **Always show the user the headline summary from this tool
`rigour_check_deep`	write	Run quality gates WITH deep LLM-powered analysis. Three-step pipeline: AST extracts facts → LLM interprets → AST verifies. Local-first by default (Qwen2.5-Coder-0.5B lite sidecar), or bring your own API key for any cloud provider.
`rigour_check_pattern`	read	CALL THIS BEFORE creating any new function, component, hook, or class. Checks if it already exists in the codebase (prevents duplication), checks for known security vulnerabilities, and BLOCKS writes to protected paths (.github/, CI/CD configs, rigour.yml). Always pass the target file path.
`rigour_context_explain`	read	Audits why specific files/services were included or excluded, cache hit/miss status, invalidation reasons, and prior agent requests.
`rigour_context_stats`	read	Returns context retrieval efficiency, candidate tokens vs returned tokens, potential avoided tokens, cache hit rate, and repeated reads prevented.
`rigour_deep_stats`	read	Get deep analysis statistics from SQLite storage. Returns recent scan scores, top issues, and score trends for a repository.
`rigour_explain`	read	Explain WHY quality gates failed in human-readable language. Use this to understand the reasoning behind each violation before fixing. For machine-readable fix instructions, use rigour_get_fix_packet instead.
`rigour_forget`	destructive	Remove a stored memory by key.
`rigour_get_config`	read	Returns the current Rigour configuration (rigour.yml) for agent reasoning.
`rigour_get_fix_packet`	read	Call this after rigour_check returns FAIL. Returns a bounded, prioritized page of violations with file locations and fix instructions. Use next_offset from the response to read further pages, then re-run rigour_check. Report only fixes that were actually verified.
`rigour_handoff`	read	Handoff task to another agent in a multi-agent workflow. Use when delegating a subtask or completing your scope. Enables verified handoff governance.
`rigour_handoff_accept`	read	Accept a pending handoff from another agent. Use to formally acknowledge receipt of a task and verify you are the intended recipient.
`rigour_hooks_check`	write	Run the fast hook checker on specific files. Same checks that run inside IDE hooks (Claude, Cursor, Cline, Windsurf). Catches: hardcoded secrets, hallucinated imports, command injection, file size. Completes in <100ms. NEW: Pass
`rigour_hooks_init`	read	Generate hook configs for AI coding tools (Claude, Cursor, Cline, Windsurf). Installs real-time quality checks and non-blocking DLP credential warnings by default. Pass dlp=false to disable DLP hooks only.
`rigour_list_gates`	read	Lists all configured quality gates and their thresholds for the current project.
`rigour_mcp_get_settings`	read	Get Rigour MCP runtime settings for this repository (.rigour/mcp-settings.json).
`rigour_mcp_set_settings`	write	Set Rigour MCP runtime settings for this repository. Currently supports deep_default_mode: off \| quick \| full.
`rigour_recall`	write	Load stored conventions. At the START of a task, call it with
`rigour_remember`	read	Store an instruction or convention to remember across sessions. Provide
`rigour_review`	read	Review the change you just made before calling it done: runs Rigour
`rigour_review_ack`	read	Record that you reviewed a function from rigour_review
`rigour_reviewer_verdict`	read	What the model reviewer last decided for this branch: the confirmed items to fix (with id, file:line, consequence and evidence), the disputed ones that are not work, which mode ran, and whether the verdict is for the current commit. Read-only: never runs a model, never dismisses.
`rigour_run`	write	Execute a command under Rigour supervision. This tool can be INTERCEPTED and ARBITRATED by the Governance Studio.
`rigour_run_supervised`	write	Run a command under FULL Supervisor Mode. Iteratively executes the command, checks quality gates, and returns fix packets until PASS or max retries reached. Use this for self-healing agent loops.
`rigour_security_audit`	read	Runs a live security audit (CVE check) on the project dependencies.
`rigour_status`	read	Quick PASS/FAIL check with JSON-friendly output for polling current project state.
`rigour_task_cost`	read	Rigour
`search`	read	Search the tracked files for a regular expression (git grep -n), optionally under one path.

Trust audit

BLOCKgrade F · trust 50/100 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	WARN
L1	Static analysis of the code	FAIL
L2	Instruction surface (what it tells the agent)	PASS
L3	Class-specific surface	WARN
L4	Behavioural (sandbox)	SKIPPED

What the source does

Filesystem: declared (7 observation(s))
Network: declared (3 observation(s))
Shell: declared (13 observation(s))
Dependencies: not all pinned
Secrets in source: found

Findings (25)

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

packages/rigour-core/src/gates/deprecated-apis-rules-lang.ts:29

api: 'pickle.load()/loads()',

Why it matters. deserialises untrusted bytes into live objects

Fix. use json or yaml.safe_load

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

packages/rigour-core/src/gates/deprecated-apis-rules-lang.ts:36

api: 'yaml.load() without Loader',

Why it matters. deserialises untrusted bytes into live objects

Fix. use json or yaml.safe_load

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

packages/rigour-core/src/gates/deprecated-apis-rules-lang.ts:38

replacement: 'yaml.safe_load() or yaml.load(data, Loader=yaml.SafeLoader)',

Why it matters. deserialises untrusted bytes into live objects

Fix. use json or yaml.safe_load

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

packages/rigour-core/src/firewall/typed-command.ts:41

/(^|\s)--eval(\s|=|$)/,

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

packages/rigour-core/src/gates/deprecated-apis-rules-lang.ts:43

api: 'exec() with user input',

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

packages/rigour-core/src/gates/deprecated-apis-rules-node.ts:177

api: 'eval()',

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

packages/rigour-core/src/deep/review-tools.ts:19

const SECRET_FILE = /(^|\/)(\.env(\..*)?|.*\.(pem|key|p12|pfx)|id_[a-z0-9]+|\.npmrc|\.netrc|credentials(\.json)?)$/i;

Why it matters. touches a credential store

HIGHHard-coded secrets · secret.db_uri · CWE-798, CWE-321

packages/rigour-cli/src/commands/demo-injections.ts:134

'    DATABASE_URL = "postgresql://admin:p@ssw0rd@prod-db:5432/app"',

MEDIUMInventory / provenance · inv.binary · CWE-1104

.pnpm-store/v11/index.db

index.db

Why it matters. a compiled or binary member cannot be reviewed from source

Fix. ship source, or explain the binary in the README

MEDIUMInformation disclosure · disclose.log_secret · CWE-209, CWE-532

packages/rigour-core/src/gates/frontend-secret-exposure.ts:146

Logger.info(`Frontend Secret Exposure Gate: scanning ${sourceFiles.length} files`);

MEDIUMNetwork egress · net.raw_ip · CWE-200, CWE-319

packages/rigour-cli/src/commands/studio.ts:536

announce(studioLaunchUrl(`http://127.0.0.1:${studioPort}`, guard));

MEDIUMNetwork egress · net.raw_ip · CWE-200, CWE-319

packages/rigour-cli/src/commands/studio.ts:566

announce(studioLaunchUrl(`http://127.0.0.1:${studioPort}`, guard));

MEDIUMNetwork egress · net.raw_ip · CWE-200, CWE-319

packages/rigour-studio/vite.config.ts:13

target: `http://127.0.0.1:${apiPort}`,

MEDIUMHard-coded secrets · secret.anthropic · CWE-798, CWE-321

packages/rigour-core/src/hooks/input-validator.test.ts:60

const result = scanInputForCredentials('sk-ant-api03-Z9Y8X7W6V5U4T3S2R1Q0');