MCP.Market

Repotector MCP server

by dir-ai·io.github.dir-ai/repotector·v1.6.0
No reviews yet

The repo guardian AI agents handshake with before they touch your code.

A85/100grade A
Write a reviewInstall
What users say
No reviews yet
Be the first
Safety scan
A85/100

full report

Adoption
New

0 stars46 downloads/wk

Reviews

Write one

Nobody has reviewed Repotector yet.

If you have run it, two minutes of your experience saves the next person an afternoon.

Repotector tools (20, 2 write)

write = sends, deletes, buys or posts

Read from the package source without running it. The installed server may list more.

  • atlas_query

    Keyword search across the repo map (paths, exports, purposes).

  • blast_radius

    Compute transitive dependents, impacted routes and components for changed files. Omit changedFiles to use git diff. Call before editing.

  • canon_check

    Check changed files against the repo canon rules. Omit changedFiles to use git diff.

  • city_map

    The info-point: DNA intent, brain pointers (atlas/genome/phenome), skeleton, and built-vs-missing so you can continue work without re-exploring.

  • claim

    Declare the paths you are working on so other agents stay out. Advisory: conflicts are surfaced, not blocked. Claims expire with your session or TTL.

  • decisions_query

    Query the deliberate choices made in this repo (what was chosen, over what, why). Consult BEFORE undoing an existing pattern or dependency.

  • declare_mission

    Bind goal + acceptance criteria + work zone to your session BEFORE working. Auto-claims your zone (conflicts surfaced), returns a one-shot briefing (gates, merge status, standing decisions, protected overlaps). depart() then reconciles the evidence against THIS contract.

  • depart

    Log your exit in the visitor register with a one-line summary of what you did. Call this when you finish.

  • dna_coverage

    Per specified clause: implemented / partial / missing / unmapped, computed from its claims vs the Atlas. The single source of built-vs-missing truth.

  • dna_diff

    Given changed files (omit to use git diff), which specified clauses they implement/modify, and which files are off-DNA (claimed by no clause). Deterministic — no semantic guessing.

  • dna_query

    Read the intent layer. Returns clauses of what was specified, marked authored (PSX Workbench mirror) or inferred (reverse-DNA, with confidence + sources) — never merged. Filter by clause id or topic.

  • find_existing

    Search the Atlas for code that already implements an intent. Call before building anything new.

  • handshake

    Front door. Declare who you are; you are logged into the visitor register and handed the ground rules, live gate, city-map summary, and a passport. Call this FIRST, before any other tool.

  • journal

    The diario di bordo: recent visits with what each agent did and left unfinished (from the visitor register). Read it to continue prior work instead of re-deriving it.

  • merge_checkwrite action

    Runs a zero-damage trial merge (git merge-tree) of HEAD against the integration base and reports clean/conflicted with the exact files, attributed to who holds each zone (claims + Merge Machine leases). Call BEFORE committing when other agents work in parallel.

  • quality_gateswrite action

    Run line-budget, structure, and secret-hygiene gates. Returns the proof verdict.

  • register

    See who has entered/left the repo and who is currently inside. The full arrival/exit log.

  • release

    Release every work-zone claim held by this session.

  • unlock

    Present the repo passphrase to unlock the gated deep-map tools for this session. Only needed if the repo is locked.

  • whats_next

    Suggests the sensible next work, derived (never invented) from DNA coverage gaps, open threads in recent departs, and TODO/FIXME markers. Every suggestion cites its evidence; an empty answer is valid.

Public scan report

scanner v0.1.9 · 2026-09-25 · same rubric, same numbers if you re-run it

no findings
  • Code scan26 source files scanned25/25
  • –Live reliabilityno gateway calls yet and no remote to proben/a
  • –Tool poisoningtools not inspected (local package is not executed); not countedn/a
  • Auth qualitylocal package, no credentials required12/15
  • Maintenancelast push 69 days ago12/15
  • Maintainer identityregistry namespace matches repository owner6/10
Overall 85/100. Components that don't apply are left out of the denominator. Any critical finding is an F.RubricAppeal a findingJSON

How Repotector compares

Against every graded MCP server, from our own scans and uptime checks. Updated after each scan run.

Safety score
85/100
higher than 70% of 48,255 servers
Tools
20
the typical server has 7
Tools that change things
10%
typical: 0%. Fewer write tools means less an agent can break
Downloads a week
46
more than 38% of 12,191 servers with a package

What its tools let an agent do: 17 read · 1 create · 2 change

Install Repotector in Claude Code, Cursor or VS Code

Runs npx -y repotector on your machine. Read the scan report first; the gateway never runs local packages.

claude mcp add repotector -- npx -y repotector
Add to Cursor

What the publisher says

From the Repotector repository's README, as published. We do not edit it. Read it on GitHub

Repotector ⬡

The repo that remembers.

Every AI agent that visits your repo starts from zero: it re-derives the map, re-discovers the conventions, re-breaks the thing the last agent just fixed. Repotector gives the repo itself a memory. An arriving agent knocks at the front door, handshakes, and is handed the map, the standing decisions, and the last visitor's briefing; on the way out it signs the register. The next agent continues in two minutes instead of wandering for an hour — and a refactor doesn't quietly break what already works.

Portable and self-contained: pure Node ESM, two dependencies (@modelcontextprotocol/sdk, zod). Works on any repo — JS/TS deeply, other stacks in honest orientation-lite.

Install into any repo

cd /path/to/your/repo
npx repotector init

init scans the repo (never fails you on day one — see grandfathered baseline below), wires the MCP server into .mcp.json, and writes the doorway blocks so every arriving agent knows to handshake. It writes only inside .repotector/ and inside REPOTECTOR:BEGIN/END markers — your prose and config are never overwritten. See SECURITY.md for the exact posture.

Under .repotector/:

Plus, at the repo root, a regenerated JOURNAL.md — the diario di bordo, newest-first, that handshake serves the tail of so the next agent continues prior work instead of re-deriving it.

The front door (handshake-first)

The MCP server refuses deep tools until an agent calls handshake — and the initialize response already tells the agent to. In return the agent gets oriented in one call, and its exit is recorded for the next one.

Shortened. The full README is on GitHub.

Nothing above is checked by us. What we check is on the safety report.

Repotector: common questions

Is Repotector MCP server safe?
Yes, by our scan: it is graded A (85/100). Read the Repotector safety report
How do I install Repotector?
It runs on your machine. Copy the Claude Code, Cursor, VS Code or Claude Desktop config from the install section.
Does Repotector need an API key?
Not as far as the registry entry and our scan can tell: no credentials are declared or required.
Is Repotector maintained?
The last commit was 109 days ago (2026-07-18). The latest release is v1.6.0.
How does Repotector compare with other MCP servers?
safety score 85/100 (higher than 70% of 48,255 servers); tools 20 (the typical server has 7); tools that change things 10% (typical: 0%. Fewer write tools means less an agent can break). Full comparison.

More from dir-ai →