Heron Logo
Heron

Built the agentBut still can’t get production access approved?

Heron interviews your AI agents, maps exactly what access they need, and generates compliance-ready reports — so security can approve with confidence

GitHub

Run locally with npx, add as a Claude Code skill, or sign in for a hosted dashboard.

See Heron in action

Watch a real audit run end-to-end — from one command to a full report with 9 systems mapped, 1 critical issue, and 5 high-severity findings.

Production rollout gets blocked when access is too broad or too vague

Most teams don’t fail because the agent is broken. They fail because the access request is too broad, too vague, or too risky for security to approve.

Security and IT do not want to approve a permission set that gives the agent more access than the task actually needs. Heron audits what the agent actually needs and generates a report that makes the access request reviewable.

01Without HeronVague wildcard scopesEndless Slackapproval threadsManual evidencecollectionCompliance leftto humans02With HeronComplete accessinventoryAutomated riskassessmentApproval-readyaudit reportsCompliance-ready docs

How Heron Works

Step 01

Give your agent one prompt

Paste this into your agent's chat or set OPENAI_BASE_URL. The agent calls Heron's endpoint and the interview starts automatically.

I need you to complete a security access review for this project.
Make sequential API calls to the Heron endpoint until it says "Interview complete".

Step 1 — Start the interview:
POST https://heron.ing/v1/chat/completions
Body: {"model":"any","messages":[{"role":"user","content":"Hi, I am ready to answer questions about this project."}]}

The response will contain a question and a session ID (in the "heron_session_id" JSON field).

Step 2 — Send each answer:
POST https://heron.ing/v1/chat/completions
Body: {"model":"any","messages":[{"role":"user","content":"YOUR ANSWER HERE"}],"heron_session_id":"SESSION_ID_FROM_STEP_1"}

Step 3 — Repeat step 2 until the response says "Interview complete".

Important: answer about THIS specific project — what you actually do, what systems
you connect to, what data you handle. Not general capabilities. Never reveal actual
secret values — just describe credential types.
Step 02

Heron interviews the agent

10 structured questions covering access, data, writes, frequency, and regulatory impact — with smart follow-ups. Typically 2-3 minutes.

Heron:

"List every system you ACTUALLY connect to in this project.
Format per system: Name → API type → Auth method
Example: Google Sheets → REST API → OAuth2 (spreadsheets.edit)

Only list systems you have actually used in this deployment — not ones that are theoretically available."

Agent:

"1. LinkedIn (via Apify) → REST API → Apify API token
2. Google Sheets → REST API (v4) → OAuth 2.0"

Heron:

"For Google Sheets, specify the exact OAuth scopes you request during the authorization process?"

Agent:

"googleapis.com/auth/spreadsheets — full read/write
googleapis.com/auth/drive.file — files created by the app"

Step 03

Get an approval-ready report

Per-system access cards, risk scoring, data quality metrics, and actionable verdict. Attach to SOC 2, ISO 27001, or GDPR evidence.

Agent Access Audit ReportMedium
Risk: MEDIUM|Systems: 2|Findings: 1 High, 1 Medium

Findings

HERON-001HIGHExcessive Google Sheets scopes
HERON-002MEDIUMPII processed without data minimization

Regulatory

EU: ReviewUS: ClearUK: Clear
APPROVE WITH CONDITIONS

Quick Start

Open source. Self-hosted. One command to start auditing your agents locally. Your data stays on your machine.

Star on GitHub
$ npx heron-ai

Then connect your agent

Paste a prompt

Copy the prompt into your agent's chat. It will call the API and complete the audit.

POST /v1/chat/completions

Override base URL

Your agent thinks it's talking to OpenAI. No code changes needed.

OPENAI_BASE_URL=...3700/v1

Scan an agent

Heron connects to the agent and conducts the interview directly.

heron scan --target ...
Claude Code Skill

/heron-audit

Install the skill once. Then type /heron-audit in any Claude Code session. Claude reads the codebase, interviews itself about what the project accesses, and generates a compliance-grade report. No server, no API keys.

  • Works in any project — Claude reads source code directly
  • Generates markdown report saved to your repo
  • Zero setup — no server, no env variables
# Install the skill
$ bash Heron/skills/heron-audit/install.sh

Then in Claude Code:

> /heron-audit

Claude reads your project and generates the audit report automatically.

Or just sign in

Same audit engine, hosted for you. No local setup, no API keys. Point your agents at the dashboard endpoint and everything is saved automatically.

Audit history

All sessions in one place

Saved reports

Full transcripts and risk analysis

Risk trends

Track compliance over time

Coming soon

Team access

Share reports across the team

Coming soon

Free. No credit card required.

Ready to audit your agents?

Free hosted dashboard. Or npx heron-ai for local use.