hardproof 0.3.0

A persistent, risk-aware engineering protocol for Hermes Agent

HARDPROOF

Software has to earn done.

Alpha software. v0.2.0 Gatehouse is the current signed public alpha on PyPI. v0.3.0 Workcells development has not begun. Commands, schemas, and contracts may change before v1.0.0.

Hardproof gives coding agents a persistent, risk-aware engineering process that turns ambiguous software requests into reviewed, verified results while preserving the evidence behind every completion claim.

A persistent, risk-aware engineering protocol for Hermes Agent.

Current Release

Release	Version	Status
Public alpha	v0.2.0 Gatehouse	Released on GitHub and PyPI
Previous	v0.1.1 Core Heat	Released on GitHub and PyPI
Previous	v0.1.0 Core Heat	Released on GitHub
Future	v0.3.0 Workcells	Not started

Install

From PyPI:

pip install hardproof
hermes plugins enable hardproof

From GitHub:

hermes plugins install asimons81/hardproof --enable

Enable

Enable the plugin in Hermes:

hermes plugins enable hardproof

Verify discovery:

hermes hardproof doctor

Quick Start

Both the CLI and slash-command surfaces work identically:

# Terminal
hermes hardproof start standard "Build API-key rotation with rollback support"
hermes hardproof status
hermes hardproof doctor

# In Hermes chat
/hardproof start standard "Build API-key rotation with rollback support"
/hardproof status

Hardproof stores state under .hardproof/, adds that directory to the repository-local Git exclude file, and injects compact stage context into bound Hermes sessions.

How Hardproof Changes an Agent Workflow

1. INT AKE — clarify the request before any code is written
2. DISCOVERY — inspect the repository, constraints, and unknowns
3. DESIGN — shape a reversible solution (requires human approval)
4. PLAN — turn the design into dependency-aware tasks (requires human approval)
5. IMPLEMENT — execute approved tasks with focused tests
6. REVIEW — challenge the implementation and risks
7. VERIFY — run checks and evaluate fresh workspace-bound evidence
8. DELIVER — prepare an evidence-backed completion handoff
9. LEARN — capture lessons or explicitly skip

Every stage has defined gates. No run completes without fresh successful evidence.

Profiles

Profile	Use Case	Key Requirements
Quick	Typo fixes, doc updates, one-line changes	Recorded skips, at least one fresh verification check
Standard	Feature work, refactors, multi-file changes	Discovery, design and plan, human design and plan approval, tracked implementation, review, fresh verification, delivery, learning decision
Critical	Auth changes, data migrations, production config	Destructive-action approvals, at least two checks, rollback and risk material, fail-closed mutation policy, human completion approval

No profile permits completion without fresh successful evidence.

Gatehouse v0.2.0 Features

Gatehouse adds strict project allow/deny/approval rules, ordered policy explanations, human-only scoped waivers, advisory risk suggestions, bounded monotonic stage graphs, configuration and migration diagnostics, and versioned Python/Node/Rust/Go policy packs. See configuration and migrations and policy packs.

Common Commands

hermes hardproof start quick|standard|critical "request"
hermes hardproof status
hermes hardproof doctor
hermes hardproof approve design|plan|completion [reason]
hermes hardproof waive <reason>
hermes hardproof pause [reason]
hermes hardproof resume [run-id]
hermes hardproof abort <reason>
hermes hardproof evidence
hermes hardproof export [path]
hermes hardproof config init|validate|explain
hermes hardproof db status|migrate [--dry-run]
hermes hardproof migrate-state
hermes hardproof runs
hermes hardproof show <run-id>
hermes hardproof policy <args>

All commands also work as /hardproof <subcommand> in Hermes chat.

Point Hermes Agent at This Repository

Hermes Agent automatically reads AGENTS.md when started in the repository. To work on Hardproof itself:

1. Clone or open the repo
2. Start Hermes from the repository root
3. Confirm AGENTS.md is loaded (Hermes prints a context-file banner on startup)
4. Follow the instructions in docs/hermes-agent-guide.md

Architecture

Hardproof is a standalone Python package discovered through the hermes_agent.plugins entry-point group. It uses only the public Hermes registration, hook, command, skill, and dispatch APIs.

1. Plugin layer — registers slash commands, CLI commands, six tools, lifecycle hooks, and nine stage skills
2. Domain layer — run profiles, stages, transitions, approval gates, and immutable event models
3. Policy layer — stage-aware mutation rules, tool policies, and human-required approval escalation
4. Storage layer — project-local SQLite database with forward-only migrations plus human-readable artifacts under .hardproof/runs/<run-id>/
5. Verification layer — workspace-bound evidence with Git HEAD capture, binary diff freshness checks, and redacted output recording

See architecture, profiles, and compatibility.

Verification Evidence

Hardproof requires fresh, workspace-bound evidence before any run can complete. Verification checks capture:

• Git HEAD commit hash at the time of verification
• Binary diff against the working tree to detect uncommitted changes
• Redacted, size-bounded output from each configured check
• Strict exit-code evaluation (only explicit zero passes)

Stale evidence — recorded against a different workspace state — is flagged and cannot satisfy completion gates.

Security Boundary

Hardproof coordinates engineering process; it is not a security sandbox. Policy hooks do not replace OS permissions, protected branches, sandboxing, isolation, or human code review.

Hardproof includes these protections:

• Force pushes and destructive Git operations are blocked
• Recognized destructive actions are blocked or require human approval
• Source mutation is stage-aware (locked in later stages)
• Human-only approval gates prevent model self-approval
• Policy decisions are recorded with cryptographic hashes of arguments and configuration

See SECURITY.md and the security model.

Privacy

Hardproof has no telemetry, no analytics, no accounts, no hosted dependencies, no remote asset fetching, and no automatic update checks. Normal local operation makes no intentional network request. Verification output is redacted and size-bounded. Policy events store argument keys and hashes rather than raw values. Nothing phones home.

Known Limitations

• Policy hooks coordinate process but are not a security sandbox or complete shell parser
• Managed runs require a Git worktree for workspace-bound freshness evidence
• Local compatibility evidence covers Hermes Agent 0.18.2 on native Windows; macOS and Linux are CI-enforced
• Downgrade from a migrated v0.2.0 database is not supported

Documentation Index

Document	Audience
docs/README.md	Full documentation index
docs/hermes-agent-guide.md	Hermes Agent operators
docs/architecture.md	Architecture overview
docs/profiles.md	Run profiles
docs/protocol.md	Protocol specification
docs/security-model.md	Security model
docs/configuration-and-migrations.md	Config and migrations
docs/policy-packs.md	Language policy packs
docs/command-reference.md	All commands and tools
docs/repository-map.md	Repository layout

Contributing

Start with CONTRIBUTING.md, GOVERNANCE.md, and CODE_OF_CONDUCT.md. Coding agents should read AGENTS.md first. Design changes use ADRs; breaking protocol changes require a public RFC issue. Contributions use the Developer Certificate of Origin rather than a CLA.

Roadmap

Version	Codename	Focus
v0.1.1	Core Heat	Standalone plugin, durable stages, SQLite state, approvals, skills, fresh evidence, reports
v0.2.0	Gatehouse	Explainable configurable policy, scoped waivers, risk suggestions, language packs
v0.3.0	Workcells	Dependency-aware task waves, resumable subagent implementers
v0.4.0	Challenge Chamber	Independent specialized reviewers, severity, fix/re-review loops
v0.5.0	Isolation	Branches, worktrees, baseline proof, rollback, backend adapters

Full roadmap: ROADMAP.md.

License

Apache-2.0. See LICENSE and NOTICE.

Affiliation

Hardproof is independent open-source software. It is not affiliated with, endorsed by, or sponsored by Hermes Agent, Nous Research, Atlassian, or any other organization.

---

Topics: hermes-agent coding-agents agentic-coding software-engineering verification developer-tools open-source python