Capabilities reference
Documented for Akmon 2.2.0.
This page is a practical map of what Akmon can do. Akmon is a producer-agnostic, tamper-evident evidence and verification layer for AI agents. The verification chain is the core capability; the bundled reference agent is one way to feed it, not the headline.
Evidence and verification (the core)
| Capability | Command | Why it matters |
|---|---|---|
| Import any agent's trace | akmon otel import <trace.json> | Brings OpenTelemetry GenAI traces (v1.37 structured and legacy v1.36-and-earlier message-event form) into AGEF; records an honest capture level |
| Content-addressed, hash-linked record | (format) | SHA-256 objects and a hash-linked event chain make the session tamper-evident by construction |
| Generate a signing key | akmon bundle keygen | Ed25519 PKCS#8 v2 key that akmon bundle sign accepts; openssl genpkey cannot produce one |
| Offline signature over the head | akmon bundle sign | Adds an Ed25519 signature over the session head, the AGEF-SIG-v1 statement |
| Operator attestation | akmon bundle attest | Records a separately signed operator-identity claim; trust attaches to the key, not the name |
| Verify integrity, signature, capture | akmon bundle verify | Checks the chain, the signature, operator attestation, and --require-capture full |
| Standalone verifier | agef-verify | Verifies a bundle with no full Akmon install, for auditors |
| Offline proof with plain openssl | akmon bundle prove-openssl | Emits statement.bin, signature.bin, pubkey.pem, and the exact openssl command; needs OpenSSL 3.x |
| Portable transport | akmon bundle export / bundle import | Moves a sealed session between machines and tools |
| Inspect, compare, redact | akmon inspect, akmon diff, akmon redact | Read a bundle, compare two sessions structurally and by field, remove sensitive bytes while preserving structure |
The standard pattern for an imported session: import a trace, export and sign the bundle, verify integrity and signature, then emit an openssl proof a stranger can check.
Capture honesty
Akmon never overstates how much it captured.
| Source | Capture level | Replay | --require-capture full |
|---|---|---|---|
| Akmon's own reference agent | full | Replays deterministically | Passes |
| OpenTelemetry import | structural | Refused | Fails (by design) |
The bundled reference agent (gold-fidelity producer)
Akmon ships a coding agent that produces full-capture sessions. It is the reference producer, not the product. Its own-agent verification surface:
akmon audit verify,akmon evidence verify, andakmon verifyfor the on-disk journal, audit chain, and evidence artifact,akmon replayfor deterministic playback,akmon slo verify --strictfor reliability thresholds.
Operating modes (reference agent)
| Mode | Command | Best use case |
|---|---|---|
| Interactive | akmon chat | supervised iterative implementation |
| Headless | akmon --yes --task "..." | CI and automation |
| JSON reporting | --output json | machine-readable orchestration |
| Plan-only | --plan | read-only scoping before edits |
| Architect | --architect | plan and implement with a model split |
| Spec workflow | akmon spec ... | structured requirements, design, and tasks |
Runtime and packaging
| Capability | Why it matters |
|---|---|
| Single Rust binary | predictable behavior across laptop, SSH host, CI runner |
Standalone verifier binary (agef-verify) | auditors verify without installing the full agent |
| Optional feature set | choose slim or full builds by environment needs |
| Terminal-first UX | works where editor plugins are unavailable |
Model and provider support (reference agent)
The bundled agent supports local and cloud providers:
- Ollama (offline and local),
- Anthropic,
- OpenAI-compatible providers,
- OpenRouter, Groq, Azure, Bedrock.
Model selection is per-task, which keeps cost and capability tuning an operator decision rather than a tooling lock-in.
Policy and safety capabilities
- permission-gated side effects,
- write diff confirmation flows,
- sandboxed filesystem boundaries,
- auditable tool and policy events.
Cost and observability capabilities
- token and cache visibility in the UI,
- cost estimates and run summaries,
- JSONL audit trail for runtime evidence.
Automation capabilities
- headless runs with budget caps,
- structured JSON run output,
- a script-friendly command model for batch operations and CI gates.
Known non-goals
- no hosted SaaS runtime (you run it),
- no mandatory IDE dependency,
- no guarantee that third-party model APIs are available,
- no certification or compliance guarantee. Akmon helps you produce evidence for frameworks like the EU AI Act, NIST AI RMF, and SOC 2; validate fit with your own legal and compliance teams.
Next steps: tutorials overview, reviewer flow, security model.