PIC Standard adds machine-verifiable contracts to agent calls and actions. Before any high-impact tool call, the agent must submit an Action Proposal (schema + verifier). If trust/evidence is insufficient, it fails closed and blocks the action.
Current version supports resolvable SHA-256 evidence. Ships with a CLI and drop-in LangGraph + MCP integrations.
Incoming cryptographic signing for trusted provenance and signed API attestations (external verification).
The PIC contract (what an agent proposes before a tool call)
PIC uses an Action Proposal JSON (schema: PIC/1.0). The agent emits it right before executing a tool:- intent: what it’s trying to do
- impact: risk class (money, privacy, compute, irreversible, ...)
- provenance: which inputs influenced the decision (and their trust)
- claims + evidence: what the agent is asserting and which evidence IDs support it
- action: the actual tool call being attempted (tool binding)
Evidence (v0.3): Resolvable SHA‑256 artifacts
PIC v0.3 adds deterministic evidence verification: evidence IDs can point to a real artifact and be validated via SHA‑256. What this gives you- evidence[].id is no longer just a label — it can be resolved to a file (file://...) and verified.
- Verification is fail‑closed: if evidence can’t be resolved or verified, high‑impact actions are blocked.
- “Trusted” becomes an output of verification (in‑memory): verified evidence IDs upgrade
- provenance[].trust to trusted before the verifier runs.
PIC Standard adds machine-verifiable contracts to AI agent tool calls for causal governance.
- for
- Developers building agentic AI systems that need provable intent and risk control.
- pricing
- open source
- license
- Apache-2.0
- v0.9.021 days ago
- v0.9.0-alpha.225 days ago
- v0.9.0-alpha.1a month ago
works withLangGraphMCPCLI
Key features
6 features of PIC Standard: Provenance & Intent Contracts
- Action Proposal schema — Agents emit a JSON contract describing intent, impact, provenance and evidence before each tool call.
- Deterministic evidence verification — SHA-256 resolvable artifacts are validated; unverifiable evidence blocks high-impact actions.
- Fail-closed security model — If verification fails, the action is automatically rejected to prevent unsafe operations.
- CLI and SDK integrations — Provides a command-line tool and drop-in integrations for LangGraph and MCP servers.
- Cryptographic signing — Supports signing of contracts and API attestations for trusted provenance.
- Risk classification — Allows agents to label impact classes such as money, privacy, compute or irreversible actions.
Use cases
- Enforcing compliance before an AI agent writes to a database.
- Blocking high-value financial transactions unless evidence is verified.
- Auditing autonomous code changes with provable provenance.
- Ensuring privacy-sensitive actions have validated data sources.
PIC Standard: Provenance & Intent Contracts vs alternatives
 PIC Standard: Provenance & Intent Contracts |  LangChain |  CrewAI |  Mastra |  Syrin - Static Contract Analysis for MCP Servers | |
|---|---|---|---|---|---|
| Best for | Verifiable contracts for AI agent actions | LLM app framework | Multi-agent orchestration | TypeScript AI agents | Static MCP contract analysis |
| Pricing | Open source | Free | Subscription | Free | Free |
| DevHunt upvotes | 3 | 0 | 0 | 0 | 24 |
| Launched | Jan 2026 | — | — | — | Feb 2026 |
- PIC Standard: Provenance & Intent Contracts vs LangChain: LangChain is a framework for building LLM apps but does not provide built-in contract verification for agent actions.
- PIC Standard: Provenance & Intent Contracts vs CrewAI: CrewAI focuses on coordinating multiple agents, lacking machine-verifiable contracts and evidence checks.
- PIC Standard: Provenance & Intent Contracts vs Mastra: Mastra is a TypeScript framework for AI agents without built-in causal governance or SHA-256 evidence verification.
- PIC Standard: Provenance & Intent Contracts vs Syrin - Static Contract Analysis for MCP Servers: Syrin offers static contract analysis for MCP servers, whereas PIC provides runtime, machine-verifiable contracts for agents.
PIC Standard: Provenance & Intent Contracts FAQ
How does PIC prevent unsafe tool calls?+
It requires agents to submit an Action Proposal; if the evidence cannot be resolved or verified, the verification fails closed and the action is blocked.
What format is the contract in?+
The contract is a JSON object following the PIC/1.0 schema, containing intent, impact, provenance, claims, evidence and the action.
Can evidence be cryptographically signed?+
Yes, upcoming versions add cryptographic signing for trusted provenance and signed API attestations.
Is there a command-line interface?+
Yes, PIC ships with a CLI for creating and validating contracts locally.
Which platforms can I integrate PIC with?+
Drop-in integrations are provided for LangGraph and MCP servers, and the library can be used in any language that can produce the JSON schema.
What happens if evidence verification fails?+
The verification is fail-closed, meaning the agent’s action is blocked and marked as untrusted.
Summarized by DevHunt from github.com · Oct 4, 2026. Details may change; check the official site.
