# VIBES Standards — Agent Manifest

You are an AI agent. Load only what you need. All URLs are absolute and stable.

## Conformance pyramid

```
                  TRACE  (IR & forensics)
              EVOLVE  (governance + learning)
          PRISM  (risk scoring)
      VERIFY  (cryptographic attestation)
  VIBES  (data substrate — REQUIRED)
```

Adopt VIBES first. Layer extensions independently as needed.

## Documents

| Standard | Spec | Conformance | Depends on |
|---|---|---|---|
| VIBES | https://itsavibe.ai/vibes-agent.md | required for all | none |
| VERIFY | https://itsavibe.ai/verify-agent.md | optional | VIBES |
| PRISM | https://itsavibe.ai/prism-agent.md | optional | VIBES |
| EVOLVE | https://itsavibe.ai/evolve-agent.md | optional | VIBES |
| TRACE | https://itsavibe.ai/trace-agent.md | optional | VIBES (+ VERIFY/PRISM recommended) |

## Machine index

https://itsavibe.ai/index.json

## File layout you must produce

```
.ai-audit/
├── config.json          # Assurance level, extension toggles
├── manifest.json        # Hash → context registry
├── annotations.jsonl    # Append-only audit log
├── audit.db             # Query cache (gitignored)
└── blobs/               # High-assurance large entries
```

## Hooks (when to write what)

- **Session start:** append `session` record; init/load `config.json`
- **Tool/prompt received:** hash context, write to manifest if novel
- **Code emitted:** append `line` or `function` annotation citing hashes
- **Delegation:** append `delegation` record (EVOLVE)
- **Decision point:** append `decision` record (EVOLVE Medium+)
- **Post-generation:** compute PRISM `risk_score` if enabled; check TRACE IoCs if enabled
- **Session end:** append session end record; produce VERIFY attestation if enabled

## Conformance signal

Tools claiming VIBES compliance MUST set:
- `config.json.standard = "VIBES"`
- `config.json.standard_version` to a published version
- `config.json.assurance_level` ∈ `{"low","medium","high"}`

Extensions opt in via:
- VERIFY: presence of signed envelopes in `.ai-audit/attestations/`
- PRISM: `config.json.risk_scoring.enabled = true`
- EVOLVE: presence of `delegation` and `decision` records
- TRACE: `config.json.incident_response.enabled = true`

## What NOT to read

- HTML pages at `/standards`, `/verify`, `/prism`, `/evolve`, `/trace` — these are human-readable narratives, not normative. Same content lives in `*-agent.md` at ~30% the token cost.
- `/u-*.html` — plain-language user-track pages. Skip these as an agent.
- `/rfc-vibes.txt` and `/standards.md` — long-form historical artifacts. Read only if `vibes-agent.md` is insufficient.

## Quick-load recipes by use case

| Goal | Load |
|---|---|
| Be VIBES-compliant | `vibes-agent.md` |
| Add signed attestation | `vibes-agent.md` + `verify-agent.md` |
| Add risk gating to CI | `vibes-agent.md` + `prism-agent.md` |
| Multi-agent governance | `vibes-agent.md` + `evolve-agent.md` |
| Detect & respond to incidents | `vibes-agent.md` + `trace-agent.md` |
| Full stack | all five `*-agent.md` (~22k tokens) |

## Section-level deep links

Every `*-agent.md` file has a Table of Contents block at the top with stable anchor links. Format: lowercased section title, spaces→hyphens, punctuation stripped. Examples:

- `https://itsavibe.ai/vibes-agent.md#2-file-layout` — just the file layout chapter
- `https://itsavibe.ai/verify-agent.md#8-trust-tiers` — just the trust tiers
- `https://itsavibe.ai/prism-agent.md#4-signal-vocabulary` — just the PRISM signal list
- `https://itsavibe.ai/evolve-agent.md#3-agent-governance` — just the governance chapter
- `https://itsavibe.ai/trace-agent.md#5-ioc-vocabulary` — just the IoC catalog

Use these when your context budget is tight or when implementing a single feature.