Signed Audit Trails Recipe
wshobson/agentsSigned audit trails give you tamper-evident proof of every action your AI agent takes. Each tool call gets a cryptographic receipt that anyone can verify later without needing the internet or trusting the operator. This skill is a cookbook for setting up that system step by step.
You start by writing a policy that controls which tools the agent can use and what commands it can run. Then every tool call is checked against this policy before it runs and signed with an Ed25519 key after it runs. The receipts are saved in a folder and can be verified with a single command.
This pattern is useful for regulated industries like finance or healthcare, for CI/CD pipelines that need proof of compliance, and for multi-party projects where partners want to verify agent behavior without trusting you.
Add Signed Audit Trails Recipe skill to your workflow
Global
mkdir -p ~/.claude/skills/signed-audit-trails-recipeProject
mkdir -p .claude/skills/signed-audit-trails-recipeSource Repository
More Security Claude Code Skills
Azure Compliancemicrosoft/azure-skills
Audit Azure compliance and Key Vault expiration with azqr scans
Firebase Security Rules Auditorfirebase/agent-skills
Find weak spots in Firestore rules and get clear fixes
Golang Securitysamber/cc-skills-golang
Find and fix security flaws in Go code using expert guidance and automated tools
Clerk Nextjs Patternsclerk/skills
Secure your Next.js app with Clerk middleware, server actions, and caching
Persona It Admingoogleworkspace/cli
Keep your digital workspace safe with IT admin security workflows
Two Factor Authentication Best Practicesbetter-auth/skills
Secure user logins with two-factor authentication using Better Auth
Gws Modelarmor Create Templategoogleworkspace/cli
Create new templates for Google Model Armor to protect AI apps from harmful content
Gws Modelarmor Sanitize Promptgoogleworkspace/cli
Clean user prompts safely with Google Model Armor templates