The passport & control plane
for agentic work.
Start with a no-secret GitHub Repository Guard, then install pre-action hooks for Claude Code, Cursor, and OpenClaw with npx. Hosted enforcement gives teams signed decisions.
Works with the agent frameworks developers use
Benchmarks
Built for real controls, not cosmetic guardrails.
Why This Exists
Prompts are suggestions.
Your agents need enforcement.
Role descriptions are soft. Quality gates should be deterministic. APort enforces what each agent can do and what they must deliver — before they proceed.
Agent shipped broken code
Prompt said 'write tests.' Agent skipped them.
Deliverable contract: 80% coverage required. Enforced.
AI changed protected repo files
Workflow, policy, or release files changed silently.
Repository Guard flags protected-path changes.
Agent ran a destructive command
Prompt injection bypassed soft safety instructions.
APort blocks destructive commands before execution.
Agent called an unapproved MCP tool
Valid credentials let the tool call proceed.
Passport policy denies the specific tool action.
Agent read a sensitive local file
Coding assistant touched .env, .ssh, or credentials.
Sensitive file-read policy blocks default-risk paths.
Task marked done, wasn't
Agent said 'done' - acceptance criteria unmet.
Deliverable policy checks required evidence before completion.
System Design
Four primitives. One system.
The same infrastructure that blocks rm -rf also checks repository changes, MCP tool calls, sensitive file reads, and GitHub merge attempts before they proceed.
Passport
Who your agent is
Open Agent Passport identity with owner, assurance level, capabilities, limits, regions, and status. Portable across runtimes and auditable by design.
Policy
What your agent can do
Pre-action authorization. Command allowlists, spending caps, file size limits, sector concentration rules. Enforced before execution, not after.
Guardrail
Where action is checked
Native hooks and GitHub checks stop unauthorized shell, file, MCP, browser, and repository actions before they proceed.
Proof
Cryptographic record of everything
Every hosted decision is signed and searchable. Teams get durable evidence for reviews, investigations, and compliance workflows.
Enforcement
Prompts change the mode.
APort changes what’s possible.
Prompt-based (honor system)
"Think like a senior architect."
"Be paranoid about security."
"Only merge if tests pass."
APort (enforced)
capabilities: [code.write, code.pr.open]
blocked: [code.push.main, code.deploy]
protected_paths: [.github/**, functions/api/verify/**]
workflow_permissions: no escalation
Implementation
60 seconds. Real guardrails.
One installer creates or selects a passport, wires the native hook, and starts recording decisions.
Repository Guard
.github/workflows/aport-guard.yml
# .github/workflows/aport-guard.yml
name: APort Repository Guard
on: [pull_request]
permissions:
contents: read
pull-requests: read
id-token: write
jobs:
guard:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: aporthq/policy-verify-action@v1
with:
mode: autoUse Cases
For developers. For enterprises.
Same primitives. Different pain points.
AI Engineering Teams
Coding agents, repo automation, and enforced quality.
- Claude Code and Cursor actions checked before shell, file, or MCP tools run
- GitHub Repository Guard flags protected paths, permission escalation, and risky PR evidence
- Agent handoffs verified with signed receipts
- Block destructive commands, secret exposure, path traversal
Regulated Industries
Independent third-party authorization with proof.
- Policy checks for finance, data, and workflow actions before execution
- Data export, file access, and MCP authorization with signed decisions
- Signed, tamper-evident attestations — not just logs
- Enterprise deployment scripts for managed developer workstations
Plans
Start free. Pay when control matters.
Free passports and report-only visibility for developers. Team and Enterprise add hosted enforcement, signed decisions, org audit, and rollout support.
Free
Passports, local guardrails, and report-only visibility.
OAP passport identity. Local passport JSON or hosted starter passport. Claude Code, Cursor, GitHub, MCP, LangChain, CrewAI, and OpenClaw quickstarts.
Start FreeTeam
Hosted enforcement and audit for one engineering team.
Hosted pre-action authorization. Ed25519-signed decisions. Organization audit and decision search.
Start Team PilotEnterprise
Org-wide rollout for security, IT, and regulated teams.
MDM-ready deployment, enforcement, and uninstall scripts. Multi-region and private deployment options by agreement. Custom policy packs.
Book Enterprise RolloutFAQ
Questions Developers Ask
Quick answers to common objections
APort is an AI agent passport and control plane for agentic work. It gives each agent an identity, enforces pre-action authorization before tools execute, and records guardrail decisions for audit.
Comparison
How APort compares.
Prompts suggest. Gateways filter. APort enforces and proves.
| Feature | APort | Prompt Rules | Runtime monitors | IAM / Vaults |
|---|---|---|---|---|
| Pre-action authorization | ✓ | — | Partial | — |
| Claude Code and Cursor hooks | ✓ | Soft | — | — |
| GitHub OIDC repository guard | ✓ | — | — | — |
| Agent identity and passport | ✓ | — | — | — |
| Cryptographic proofs | Ed25519 | — | — | — |
| Tamper-evident attestations | ✓ | — | — | — |
| MCP tool authorization | ✓ | — | — | — |
| Protected path evidence | ✓ | — | — | — |
| Framework agnostic | ✓ | ✓ | Varies | — |
| Open source | ✓ | Varies | — | — |
APort complements IAM, vaults, sandboxes, and runtime monitors: they manage identity and environments; APort checks whether the specific agent action is authorized now.
Start with one agent or one repo
Install guardrails today.
Start free with a hosted passport, protect a GitHub repository, then upgrade when your team needs org-wide policy, audit, and managed rollout.
