Evaluate the repository
The assessment reflects a repository snapshot. Passed and failed checks contribute to scores; checks that cannot be assessed are skipped.
Agent Readiness · part of Developer Intelligence
Find what is holding your coding agents back
Larridin assesses whether coding agents can understand, change, and verify your repository. See gaps in documentation, tooling, tests, and automation, then prioritize what to fix next.
How ready is this repository for agents?
Example scan · main
L3
17 checks from L4
L1Baseline
L2Documented
L3Agent-ready
L4Optimized
L5Autonomous
Next steps · to reach L4
17 checks to go
1AGENTS.md Structure DocumentationBreak the agent guide into clearly headed sections (Commands, Architecture, Conventions).
2Automated Security Review SecurityEnable CodeQL or Semgrep on every pull request.
3Build Performance Tracking Build SystemEnable remote build caching and track build times in CI.
4Runbooks Documented ObservabilityWrite runbooks for the top operational scenarios and link them from alerts.
Trusted by AI-forward enterprises
Actionable Readiness Scores
Every failing check includes the evidence behind it, why it matters to an agent, and the shortest path to passing. Larridin prioritizes foundational gaps first, so teams fix the problems creating the most friction
Skipped, never failed · a check the scan cannot assess is excluded from the score, and a level cannot clear on a sparse sample either
Contiguous levels · a repo cannot reach Level 4 over an uncleared Level 3, however many higher checks pass
A grade per category · nine categories, A to F, so the radar shows where the gap is before you open the list
Explore the scoring categories
larridin/larridin-workflow-intelligence
L3
Agent-ready
L1Baseline
L2Documented
L3Agent-ready
L4Optimized
L5Autonomous
Evaluated from the repository at a pinned commit. Unassessable checks are skipped and excluded from every score.
L1
Baseline
The starting tier. Repos climb from here as the basic checks (formatter, linter, types, unit tests) come online and run locally, so agents can change code and know it compiles.
L2
Documented
The workflows are written down: agent guides, environment templates, schema, ownership. Agents stop guessing at tribal knowledge.
L3
Agent-ready
Standards are enforced through automation: CI, hooks, tracing, release workflows. Agents get automated feedback before a change is reviewed and merged.
L4
Optimized
Feedback loops are fast and measured: coverage gates, flaky-test tracking, build timing. Agents iterate at machine speed instead of CI speed.
L5
Autonomous
Production signal feeds work back into the loop: canary rollouts, auto-rollback, error-to-issue pipelines. The system improves itself with agents in it.
Do this next
Every failing check comes with fix guidance.
The report lists the checks that block the next level and names the tool or file to change for each one: Husky, CodeQL, gitleaks, a coverage threshold. An engineer, or the agent, can start without a meeting.
Check, why, fix · what the scan looks for, why it matters to an agent, and the shortest path to passing
Foundational gaps first · clear lower-level gaps before progressing to the next readiness level
What's working, too · strengths the repo already demonstrates, with the evidence the scan found
Do this next
1AGENTS.md StructureBreak the agent guide into clearly headed sections (Commands, Architecture, Conventions).DocumentationL4
2Automated Security ReviewEnable CodeQL or Semgrep on every pull request.SecurityL4
3Build Performance TrackingEnable remote build caching and track build times in CI.Build SystemL4
4Runbooks DocumentedWrite runbooks for the top operational scenarios and link them from alerts.ObservabilityL4
5Secret ScanningEnable push protection and run gitleaks in CI.SecurityL4
What's workingmost advanced passes
Test Isolation · L3
Distributed Tracing · L3
Integration Tests Exist · L3
Pre-commit Hooks · L3
Example failing checks. A level clears when at least half its evaluated checks pass and at least half its checks were evaluated, with every lower level cleared.
DocumentationL4Failing
AGENTS.md Structure
Evidence · AGENTS.md found at the repo root · 312 lines · 1 heading · last touched 41 days ago
Example fix workflow
Use with your coding agentBring the guidance into your editor
What we look forA structured agent guide with clearly delineated sections (multiple headings).
Why it mattersA sectioned guide lets agents jump straight to the part they need instead of re-reading one wall of text.
How to fixBreak the agent guide into clearly headed sections (Commands, Architecture, Conventions).
Suggested sections, built from the evidence
1Commands: build, run, check and the one test command (pnpm test)
2Architecture: repo layout and the main request flow
3Conventions: naming rules and test layout, already followed in src/
4Git workflow: gh operations, branch protection, PR template
Example reviewed PR · 1 file changed · verify on the next scan
Passing
Example workflow: give the guidance to your coding agent, review the change in a normal pull request, then check the next scan.
Fix guidance
Fix it with the agent you already use.
Use a failing check's evidence and suggested fix as a brief for your coding agent. Review the resulting change through your normal pull request process, then use the next scan to see whether the check passes.
A clear target · the check explains what the scan looks for and what needs to change
Your agent, your account · bring the guidance to Claude Code, Codex, or the agent in your existing workflow
Evidence to work from · use the evidence and suggested fix to give the agent a specific task
Reviewed as usual · ask your agent for a change you can test and review before merging
Agent readiness, checked from every angle
Each category says what the mechanism is and what it does for an agent.
Style & Validation
Formatters, linters, and type checkers catch mistakes in seconds, locally. Agents self-correct before a human ever sees the diff.
Build System
Deterministic builds and documented commands let agents verify their own work. Fast, automated pipelines turn every change into a tight edit-verify loop.
Testing
Tests are the strongest signal an agent can check its work against. Runnable, isolated, trustworthy suites make agent changes safe to merge.
Documentation
Written instructions capture what "everyone just knows." Agents can't absorb tribal knowledge, so the docs are their onboarding.
Dev Environment
Reproducible environments boot with one command. When agents and developers work in identical environments, setup failures disappear.
Debugging & Observability
Structured logs, traces, and metrics show what the code actually does at runtime. Agents debug from evidence instead of guesswork.
Security
Guardrails like branch protection, secret scanning, and ownership keep agent mistakes from becoming incidents. Agents move fast inside hard limits.
Task Discovery
Well-structured issues and templates make the backlog machine-readable. Agents can find, scope, and pick up work on their own.
Product & Analytics
Usage analytics and error pipelines turn production signal into prioritized work. Agents see impact, not just code.
Free for public repos
Scan any public repository at code.larridin.com.
Larridin Code reads the merged-PR and commit history of public GitHub repositories and scores them with the same 84 checks, beside the engineering signals behind the score: work distribution, churn, review health, cycle time.
106
repositories scored
47,450
merged PRs read
93,323
commits analyzed
Larridin Code
OverviewContributorsTrendsRiskPull requestsMethodology
RepositoryLevelChecksMerged PRs
BerriAI/litellmL5Autonomous73%473
apache/supersetL5Autonomous68%591
openclaw/openclawL5Autonomous68%229
Agent readiness and engineering signals for a curated set of public GitHub repositories. Read-only, from public data.
All 106 repositoriesAcross the org
One number for every repository you own.
The org view rolls every active repo into a level distribution, the share at Agent-ready or above, and the checks most repositories fail. Platform teams can identify common gaps and coordinate fixes across repositories.
Activity-weighted · repos count by recent merged PRs, so a dormant one cannot drag the average
Recent activity in focus · compare the repositories active in the reporting window
Applicable checks · checks the scan cannot assess are excluded from scoring
Agent Readiness · Organization
L2.4
average level, activity-weighted
15 of 35
repos at Agent-ready or above
4
checks failing in ten or more repos
Level distribution
L16
L214
L312
L43
L50
Most-failed checksrepos failing
AGENTS.md L2Add AGENTS.md with the commands, architecture map, and conventions agents must follow.23
Pre-commit Hooks L3Install Husky or pre-commit and run the formatter, linter, and type check on staged files.19
Integration Tests Exist L3Stand up an integration suite covering the top user flows.17
Branch Protection L2Protect main: require PRs, passing checks, and at least one review.14
Example rollup of repositories active in the last 90 days, weighted by merged PRs.
A closer look
How repository readiness is assessed
The assessment uses 84 binary checks across nine categories and five levels. Reports connect a failing check to the repository evidence and guidance for addressing it.
Clear levels in order
A level clears when at least half its checks were evaluated and at least half the evaluated checks pass. Every preceding level must also clear.
Fix and reassess
Use the check’s guidance to make a reviewable change, run the relevant verification, and reassess the repository. A readiness level is not a guarantee that an agent can ship safely.
Questions about Agent Readiness
Is this an assessment of our codebase or our organization?
Agent Readiness assesses a repository’s environment, documentation, tooling, and feedback loops. Organizational AI maturity includes people and operating practices and is a separate assessment.
Do skipped checks count as failures?
No. Skipped checks are excluded from the pass rate. A level still needs sufficient evaluated coverage, so a repository cannot clear a level from only a small sample.
Does Larridin automatically fix every failed check?
The report provides evidence and fix guidance. You can implement a change yourself or hand it to a coding agent, review it through your normal pull-request process, and run another assessment. The assessment does not guarantee an automatic fix for every check.
How should I interpret the product examples?
The product views on this page use illustrative data to explain the metrics and workflows. For metric definitions, sample calculations, and assumptions, see our measurement methodology.
Find out before the first agent session, not after the hundredth.
Start with one public repository, free. Then connect your org and grade every repository this week.