grith
OS-level agent supervisor

Your agent decides what's safe. grith decides what runs.

Claude, Codex, Aider and friends approve their own actions with probabilistic models. grith sits underneath the agent and intercepts every syscall - scoring it against 18 filters before it touches your machine.

grith supervisorintercept #4821
fs.read0.02ms
path /home/dev/.ssh/id_rsa
agent claude-code · auto-approve
18 filters
  • sensitive-path-heuristic, contribution +5.0, high
  • secret-scan, contribution +4.5, high
  • path-match, contribution +2.0, medium
  • taint, contribution +1.0, medium
  • operation-risk, contribution +0.5, low
⛔ DENIEDscore 13.0 · deny > 8.0
SupervisesClaudeCodexAiderClineGoose
The gap

A malicious README, a poisoned dependency, one clever prompt injection - and the agent is reading your SSH keys, with you out of the loop.

The fix

grith puts a deterministic supervisor between the agent and the OS. Every call is scored. Ambiguous ones wait for a human.

One line, any agent

Wrap the agent. Keep the loop.

grith hooks the OS, not the app - so there's nothing to integrate. Start any agent through grith and every call it makes is intercepted, scored, and either allowed, queued, or denied. Auto-approve stops meaning auto-trust.

$ grith exec -- claude-code
grith ● 1,208 calls intercepted
1,204 calls allowed
3 queued for review
1 denied secret-scan
01

Below the agent, not beside it

grith intercepts at the syscall boundary. The agent can't route around a supervisor it doesn't know is there.

02

18 filters, every call

Secret scanning, sensitive-path detection, egress policy, destructive-op checks, taint tracking - scored in milliseconds.

03

Human review queue

Ambiguous calls pause and route to a reviewer. The agent waits. You decide. Nothing runs on a maybe.

04

Agent-agnostic

One supervisor for Claude, Codex, Aider, Cline, Goose and anything else that shells out.

The system of record

Blocking is the start. The record is the product.

Every intercepted call is scored, decided, and written down. That turns agent activity into something you can measure, replay, and hand to an auditor.

fleet · last 24hexample data14 agents
48,210calls supervised
  • allowed45,164
  • queued2,731
  • denied315
top filtersecret-scan · 1,204×

Analytics

Allow / queue / deny rates, top filters triggered, and per-agent risk across your whole fleet.

Audit trails

A signed, immutable log of every call and verdict - replayable, exportable, and tamper-evident.

Compliance-ready

Policy-as-code mapped to controls. Export SOC 2-aligned evidence straight from the record.

Enterprise controls

SSO / SAML (planned), RBAC (planned), centrally-managed policies, fleet rollout, and self-hosted deployment (planned).

Built forAudit evidence exportSOC 2-aligned reportingSigned audit trailSSO / SAML (planned)RBAC (planned)Self-hosted (planned)

Stop trusting. Start supervising.

Latest writing

Technical work on AI agent security, syscall interception, and the trust boundary.

Read all posts