AVAILABLE NOW
Guardrails that live in your repository
Install policies into your own repository: committed with the code, enforced at commit and in CI for contributors who run chock sync, and for agents whose hooks are set up.
What each policy does in a git repository
35
At commit
Chock refuses the commit, locally and in CI.
11
In agent
The agent's hooks check this once chock sync sets them up. If the hook crashes, the action goes through.
25
Advisory
The agent reads the text, and nothing stops a violation.
Each build counts these from the catalog.
How it works
Pick your policies
Answer three questions, or take a preset. The builder writes one block of commands.
Run the commands in your repo
They add each policy at a pinned commit and hash, wire the hooks and write files. You commit those files with your code.
Every clone runs one command
Git never clones hooks, so each person who clones the repository runs
chock sync --repo .once.
Install chock first
The commands need chock on your machine. Read the command before you run it.
pipx install chock
The Install page has the details and the other ways to install. →
Every policy, by what it does in a repository
Each name opens the policy's page. This page reads each tier from the catalog. A policy's grade on one agent does not change its tier.
Checks at commit (35)
- Agentic Code Security
agentic-code-security - Block Destructive Commands
block-destructive-commands - Block Files That Skip Git Hooks
block-hook-bypass-in-files - Block Invisible Unicode
block-invisible-unicode - Block Known-Malicious Packages
compromised-package-ioc - Block Risky Characters in File and Branch Names
refname-filename-metachar - Block Test Skips
block-test-skips - Block Unpinned Agent Components
block-unpinned-agent-components - Block Unsafe Code Execution
block-unsafe-code-execution - Block Unsafe Package Registry Settings
registry-config - Block Weakened Build Hardening
hardening-flags - Block Wildcard Agent Permissions
block-wildcard-agent-permissions - Block Wildcard IAM Grants
block-wildcard-iam - CI GitHub Actions Security
ci-github-actions-security - Dependency Allowlist
verify-dependency-exists - Dockerfile and Compose Security
dockerfile-compose-security - Guard Deletion and Mitigation Removal
guard-deletion - Guard Memory Writes
guard-memory-writes - IAM Policy Scan
iam-policy-scan - Java Security Rules
java-security - Keep Accessibility From Regressing
no-a11y-regression - Keep Changes Small Enough to Review
limit-diff-size - Lockfile Integrity
lockfile-integrity - Pin GitHub Actions
pin-github-actions - Protect Commit Privacy
protect-commit-privacy - Protect Main Branch
protect-main-branch - Protect Test Integrity
protect-test-integrity - Review Added Scanner Suppressions
scan-suppression-markers - Review Scripts That Run on Install
package-lifecycle-scripts - Review Values That Look Like Secrets
scan-secrets-entropy - Scan Agent Permission Settings
agent-permissions-scan - Scan Instruction Files
scan-instruction-files - Scan Secret Files
scan-secret-files - Scan Secrets
scan-secrets - Verify MCP Allowlist
verify-mcp-allowlist
Checks in the agent (11)
- Block Agents Started With Safety Off
block-unguarded-agent-spawn - Block Curl Piped to Shell
block-curl-pipe-sh - Block Persistence and Publishing Commands
block-persistence-shapes - Block Reading Stored Credentials
block-secret-store-reads - Block Skipping Git Hooks
block-no-verify - Block Unapproved Egress
block-unapproved-egress - Protect Agent Config
protect-agent-config - Protect CI Workflows
protect-ci-workflows - Review Cloud Identity Switches
ask-credential-switching - Token Efficiency
token-efficiency - Use Firecrawl Only as a Fallback
firecrawl-fallback-only
Advisory (25)
- Agent Discipline
agent-discipline - Code Safety
code-safety - Context Hygiene
context-hygiene - EU AI Act — High-Risk Triage
eu-ai-act-high-risk-triage - EU AI Act — Prohibited Practices
eu-ai-act-prohibited-practices - EU AI Act — Transparency
eu-ai-act-transparency - Flag Dev Tools That Run Code on Open
agent-devenv-autoexec - Flag Downloads Run as Code in Files
block-fetch-exec-in-files - Flag Hidden Binary Files in Tests
opaque-blob-guard - Git Safety
git-safety - Injection Defense
injection-defense - Match the Owner's Coding Style
chock-mise - Memory Discipline
memory-discipline - OWASP ASI01 — Agent Goal Hijack
owasp-asi01-agent-goal-hijack - OWASP ASI02 — Tool Misuse & Exploitation
owasp-asi02-tool-misuse - OWASP ASI03 — Identity & Privilege Abuse
owasp-asi03-identity-privilege-abuse - OWASP ASI04 — Agentic Supply Chain Vulnerabilities
owasp-asi04-agentic-supply-chain - OWASP ASI05 — Unexpected Code Execution
owasp-asi05-unexpected-code-execution - OWASP ASI06 — Memory & Context Poisoning
owasp-asi06-memory-context-poisoning - OWASP ASI07 — Insecure Inter-Agent Communication
owasp-asi07-insecure-inter-agent-communication - OWASP ASI08 — Cascading Failures
owasp-asi08-cascading-failures - OWASP ASI09 — Human-Agent Trust Exploitation
owasp-asi09-human-agent-trust - OWASP ASI10 — Rogue Agents
owasp-asi10-rogue-agents - Review Like a Red Team
review-like-a-red-team - Scan Hidden Content
scan-hidden-content