Skip to content
How it learnsFrom coding sessions, reviews, Slack and docs.Only what fitsEach session gets the few rules its task needs.Pull request reviewEvery PR checked against your team’s rules.Keep it trueA gate, conflicts, disputes and stale rules, handled.HandoffsCarry the work to another session or a teammate.SecurityWhat it sends, what it changes and how to leave.
Engineering teamsOne agent or several: what one person learns, every session uses.PR review onlyInstall the GitHub App. Nothing on anyone’s machine.Teams that mix agentsClaude Code, Cursor, Codex and Copilot, the same rules.On your ownYour corrections follow you across repos, machines and agents.Book a call30 minutes with Roei, if you want a hand. Never required.
DocsPricingSign in
Connect your agents →
For teamsOn your ownPR review onlyHow it learnsOnly what fitsPull request reviewKeep it trueSecurityDocsPricingSign inConnect your agents →
Home/Blog

Blog

AGENTS.md, Cursor rules and Copilot instructions

  • AGENTS.md vs CLAUDE.md
  • 3,905 files checked
  • Sync rules across repos
  • Why Copilot ignores it
  • Cursor Team Rules
  • Why Codex ignores it
  • Cursor rules
  • copilot-instructions.md
  • Claude Code and AGENTS.md
  • AGENTS.md research

CLAUDE.md and context engineering

  • Why Claude ignores it
  • Context rot, stale rules
  • Growing CLAUDE.md
  • Which rules got used
  • CLAUDE.md length
  • Claude Code best practices
  • Context engineering, Claude 5
  • Karpathy CLAUDE.md
  • Committing CLAUDE.md

Built-in agent memory

  • Agent memory compared
  • Make Claude Code remember
  • GitHub Copilot Memory
  • Claude Tag vs Claude Code
  • Codex memories
  • Cursor memories
  • Claude Code auto memory
  • Memory or documentation?

Team knowledge and shared conventions

  • Claude Code team memory
  • Shared memory across agents
  • Hand off between agents
  • Claude Code second brain
  • Company brain
  • Slack decisions
  • Memory poisoning
  • Karpathy's LLM wiki
  • ADRs for agents

Skills, hooks and permissions

  • Rules vs skills
  • Claude Code hooks
  • Codex skills
  • Auto mode
  • CLAUDE.md vs skills vs hooks
  • Claude Code skills

AI code review and PR guardrails

  • PR review against AGENTS.md
  • Review comments to rules

Memory tools and alternatives

  • Best company brain tools
  • Unblocked alternatives
  • claude-mem alternatives
  • Supermemory alternatives
  • Supermemory Company Brain alternatives
  • Mem0 alternatives
  • Best MCP memory servers
  • claude-mem

Blog

Unblocked alternatives for coding agents (2026)

Published November 15, 2026

We make Harbor, one of the alternatives below. Everything here about Unblocked and the other tools is from their own pages, read on 15 November 2026.

The short answer depends on why you are leaving. If you want the rules in the session before the agent asks, with a gate on what gets in and a count of what was used, Harbor. If you want one memory for the whole company, not just engineering, Sentra or Coworker. If your rules should live as files in every repo, Packmind. For memory as files a small team can read, ByteRover. If you mostly need search across company apps, Glean.

If you are leaving Unblocked becauseLook at
The agent only gets context when it thinks to askHarbor
Nobody reviews what it learnsHarbor, Packmind, ByteRover
You cannot see which context an agent usedHarbor
You need it beyond engineeringSentra, Coworker, Glean
There is no free planHarbor, ByteRover, Packmind
You want to host it yourselfPackmind, Sentra (VPC or on-prem)

What Unblocked does for coding agents

Unblocked is a context engine for engineering teams. It reads code and pull requests, Slack and Teams, docs, issue trackers and CI, and answers people and agents with citations. For coding agents it runs an MCP server with two read-only tools: one that researches a question and returns a cited answer, and one that fetches what sits behind a link. It names Claude Code, Cursor, Copilot in VS Code, Codex, JetBrains, OpenCode and Claude Desktop.

It is strong at what it does: answers on demand from everything your engineers have written, with nobody curating rules, SOC 2 Type II, and on Enterprise, answers limited to what each person can see. The Platform plan also includes Unblocked Code, an agent that opens draft pull requests. Our Harbor vs Unblocked page has the full table.

Why teams look for an alternative

These are the limits Unblocked's own pages make visible:

  • The agent has to ask. Context arrives when the agent calls a tool. The MCP page describes no hooks and nothing added at the start of a session, and says Unblocked complements AGENTS.md and CLAUDE.md rather than replacing them.
  • Nobody approves what it learns. It reconciles sources on its own, server side. No review step is described for what becomes a shared answer.
  • You see tool calls, not what was used. Admins see MCP calls by agent and answer ratings, not which piece of context an agent acted on.
  • There is no free plan. $29 a user a month billed yearly, or $35 monthly, after a 21-day trial.

1. Harbor: the team's decisions, before the agent asks

Harbor learns what your team decided from Slack, pull request reviews and docs, plus what agent sessions record. Hooks give each Claude Code, Cursor, Codex or Copilot session the rules that apply to its task before the agent writes a line, and the same rules are on an MCP server. You set the gate: Manual holds everything for a person, Hybrid (the default) adds routine rules and holds the important ones, and in every mode a secret, an install step or a dangerous command waits for a person.

  • Instead of Unblocked because: the rules are in the session from the first turn, nothing reaches every agent without passing your gate, and every rule shows how often it was served and how often an answer cited it (which rules got used). Block rules stop a risky command before it runs.
  • Stay on Unblocked if: you want answers synthesised on demand from everything, with nobody deciding what counts as a rule, or you need SSO, SCIM and SOC 2 today.
  • Watch out for: it is new. Public sign-up opened on 31 October 2026. It is hosted only and wires four agents.
  • Price: Free costs nothing for up to 2 people, on one source. Pro is $24 per person a month billed yearly.

2. Sentra: one memory for the whole company

Sentra reads more than 200 company systems, from email, chat and meetings to the CRM, ERP and HR, and keeps track of what is true now and why it changed. Coding agents read it over MCP.

  • Instead of Unblocked because: it covers every department, not only engineering, and deploys in its cloud, your VPC or air-gapped on premises.
  • Watch out for: its approval step is for what it drafts or sends, not for what enters its memory, and nothing on its pages is specific to code.
  • Price: not published. You book a demo.
  • Side by side: Harbor vs Sentra.

3. Coworker: a context layer for every AI tool

Coworker connects more than 50 work tools into one company graph and sells its own apps on top, including a coding agent for the terminal. Any MCP client can read it, and its pricing page names Claude Code and Cursor.

  • Instead of Unblocked because: one layer for every AI tool in the company, with SOC 2 Type II, GDPR and CASA Tier 2.
  • Watch out for: one Enterprise plan, after a demo. It does not describe how context reaches a session, or any review of what agents are told.
  • Price: not published.

4. Packmind: one playbook, written into every repo

Packmind keeps one playbook of standards, commands and skills and writes it into each agent's own files across your repos, from .claude/rules to AGENTS.md.

  • Instead of Unblocked because: the rules are files in the repo, open source under Apache 2.0, and self-hostable. Change Proposals let the team review an edit before it applies, on Enterprise.
  • Watch out for: rules change only when someone proposes a change. It does not read Slack or pull requests inside the app.
  • Price: the open-source edition is free. Enterprise is priced on request.

5. ByteRover: memory as files a small team shares

A desktop app and an agent skill save decisions, fixes and rules as context files in spaces you share with teammates. Through its company brain plugin, what an agent learns waits as a proposal until someone reviews it.

  • Instead of Unblocked because: a review step, memory you can read and edit, ten agents named, a free plan and SOC 2 Type II.
  • Watch out for: it does not read Slack, pull requests or docs, and recall happens when the agent is asked to check.
  • Price: free, then Pro at $15 a month billed yearly or $19 monthly.
  • Side by side: Harbor vs ByteRover.

6. Glean: search across the company's apps

Glean is enterprise search, and its managed MCP server opens that index to Cursor, Claude Code, VS Code and Windsurf, with a plugin for Codex. It is the pick when the question is “where is that doc” rather than “what did we decide”. Its engineering guide describes no review step for what gets indexed. Pricing is by demo.

Side by side

Reaches the agentWho decides what gets inShows what was usedFree plan
UnblockedWhen the agent calls a toolAutomaticTool calls per agentNo, 21-day trial
HarborRules that apply, at session start, plus MCPA gate you setServed and cited, per ruleYes, 2 people
SentraWhen the agent calls its MCP serverA person approves what it sends, not what it storesNot documentedNo, demo
CoworkerAny MCP clientNot documentedNot documentedNo, demo
PackmindRule files in the repoDirect edits, or review on EnterpriseWhich version is whereOpen source
ByteRoverWhen the agent checksA reviewer, for its company brain pluginSources on each recallYes
GleanWhen the agent calls its MCP serverNot describedNot describedNo, demo

The full field, nine tools with prices, is in the best company brain tools for coding agents.

Try Harbor next to Unblocked

The two can run side by side: Unblocked answers when the agent asks, Harbor puts the team's rules in the session before it does. Connect one source, run a week, and look at which rules were served and cited. Start free.

Questions

What is the best alternative to Unblocked for coding agents?

It depends on why you are leaving. For the rules in the session before the agent asks, with a gate on what gets in and a count of what was used, Harbor. For one memory across the whole company, Sentra or Coworker. For rules as files in every repo, Packmind. For memory files a small team can read and review, ByteRover. For search across company apps, Glean.

Does Unblocked add context at the start of a Claude Code session?

Its MCP page, read 15 November 2026, describes two read-only tools the agent calls: one researches a question and returns a cited answer, one fetches what is behind a link. It describes no hooks or session-start context, and says Unblocked complements AGENTS.md and CLAUDE.md rather than replacing them.

Does Unblocked have a free plan?

No. Its pricing page lists Platform at $29 a user a month billed yearly or $35 monthly, after a 21-day trial, and Enterprise on request.

Is Harbor an Unblocked alternative?

For a related job. Unblocked answers questions on demand from everything your engineers wrote. Harbor learns what your team decided from Slack, pull request reviews and docs, puts the rules that apply into every Claude Code, Cursor, Codex or Copilot session at the start, through a gate you set, and counts per rule how often each was served and cited. Harbor's Free plan covers 2 people.

More on Memory tools and alternatives

  • The best company brain tools for coding agents (2026)
  • claude-mem alternatives for Codex, Cursor and teams
  • Supermemory alternatives for coding agents (2026)
  • Supermemory Company Brain alternatives (2026)

Harbor compared with each tool →

Your team’s rules, in every agent, checked on every PR.

Harbor is shared memory for your team’s Claude Code, Cursor, Codex and Copilot. Free for you and one teammate. No card. No end date.

Start free →

Product

  • How it learns
  • Only what fits
  • Pull request review
  • Keep it true
  • Handoffs
  • The result
  • Security
  • For coding agents
  • Pricing

Developers

  • Docs
  • Quickstart
  • Install
  • CLI
  • MCP server
  • Served and cited
  • Rule-file checker
  • Blog

Guides

  • Why Claude ignores CLAUDE.md
  • AGENTS.md vs CLAUDE.md
  • Share Claude Code memory with a team
  • Why Codex ignores AGENTS.md
  • Why Cursor ignores rules
  • Copilot instructions vs AGENTS.md
  • Shared memory across agents
  • Make Claude Code remember

Compare

  • All comparisons
  • vs Mem0
  • vs Zep
  • vs Unblocked
  • vs ByteRover
  • vs Cognee
  • vs Hindsight
  • vs OpenViking
  • vs Supermemory
  • vs Glen
  • vs MemPalace
  • vs Sentra

Company

  • Questions
  • Contact
  • X
  • LinkedIn
  • Privacy Policy
  • Terms of Service
  • Data Processing Agreement
  • Refund Policy
© 2026 HarborConnect your agents