Product
The hosted dashboard for instruction quality scores, drift detection, and rollouts.
After you run the scanner, hosted DirectiveOps is where platform teams track quality scores, detect drift, and — on Team — roll out fixes across the fleet with reviewable PRs.
Score instruction file quality. Detect drift and conflicts across AGENTS.md, CLAUDE.md, and Copilot files. Fix what's costing you tokens and degrading agent output.
DirectiveOps scans your instruction files, scores their quality, detects conflicts across tools, and shows you exactly what to fix. Teams upgrade to roll out standards and track drift with reviewable PRs.
How it works
Inside the hosted dashboard in four steps
Open the hosted dashboard to see which repositories have AGENTS.md, CLAUDE.md, GEMINI.md, and Copilot instruction files — and which ones drift from the standard.
Drill into repositories to review normalized constitutions, drift findings, and risky imports before you change anything.
Use org-level templates and policy rules to define how instruction files should look across repos, including allowed overrides.
Preview the changes, create rollout batches, open PRs across repositories, and track acceptance and remaining drift.
OSS scanner
Run the scanner locally, then plug results into hosted.
Run the free CLI against any repository checkout to discover instruction files and findings — no account required. When you are ready for fleet-wide standards and rollouts, connect GitHub so hosted DirectiveOps can turn that data into repo inventory, drift views, and rollout candidates across your fleet.
CLI scanner commands
From inside a repo, use --path . to scan the current directory.
Need fleet-wide standards and rollout? See pricing →
Already scanned? Connect GitHub to see your fleet.
Hosted product
Fleet-wide standards and rollout when the team is ready
Once you know what needs to change, hosted DirectiveOps adds central repo inventory, shared standards, and rollout PRs you can track across the fleet.
Hosted
Quality dashboard
Org-wide quality scores, per-repo breakdowns, score trends, and estimated token cost impact.
Deep dives
Instruction viewer
Review layers, precedence, imports, and instructions before approving changes.
Risk & drift
Drift detection
Surface conflicts, stale references, and cross-tool inconsistencies so you can fix drift or approve exceptions.
Rollouts & PRs
Rollout engine
Define standards, preview changes, open PRs, and track rollout status across repositories (Team+).
Precedence
See which directive layers apply in each repo
Repository-wide files, path-scoped Copilot instructions, and Codex-style AGENTS.override.md all contribute to effective agent behavior. DirectiveOps surfaces layers and precedence so platform teams can review what agents are told to follow—without executing agents for you.
CI integration
Add to CI in 2 minutes
Run the scanner on every push. Catch instruction drift before it ships.
GitHub Actions
name: directiveops-scan
on:
pull_request:
paths:
- 'AGENTS.md'
- 'CLAUDE.md'
- '.github/**'
jobs:
scan:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: actions/setup-node@v4
with:
node-version: '20'
- run: npm ci
- name: Run DirectiveOps scanner
run: |
npx directiveops-scanner scan --path . --format json --quiet --fail-on highThis workflow runs the DirectiveOps scanner on every pull request that touches instruction files. It fails the build when high-severity drift findings are detected.
Ready for hosted
Try the dashboard, then sign in and choose a plan.
You've seen the scanner. See the hosted dashboard, pick a plan, and roll out standards across your fleet.
FAQ
Common questions about AI coding instruction governance
How do you roll out AGENTS.md across many repositories?
Run the DirectiveOps scanner to discover existing AGENTS.md files, define an org-level template in the hosted product, preview the changes it would make, and generate rollout pull requests for selected repos. You can then track acceptance, rejection, and remaining drift from the dashboard.
What causes instruction drift in AI coding assistants?
Instruction drift happens when AGENTS.md, CLAUDE.md, GEMINI.md, and Copilot instruction files are edited locally in each repo without a shared standard. Teams copy, tweak, or forget to update files, and there is no central view of how instructions diverge over time.
How do teams standardize coding instructions for AI agents?
Teams use DirectiveOps to define org-level instruction templates, compare existing files against those templates, and enforce changes via rollout pull requests rather than ad-hoc edits. The findings engine highlights conflicts and missing directives so reviewers can adjust the standard or approve exceptions.
Can we keep the scanner local while using hosted for rollouts?
Yes. The OSS scanner remains a local tool for discovery and basic findings. When you are ready for fleet-wide standards and rollouts, connect GitHub to the hosted product to sync repositories, track drift, and create PR-based updates while still running local scans where you need them.