history
An agent's work history — its transparency log filtered to signed session.v1 records, every envelope re-verified on your machine.
treeship history lists an agent's signed session.v1 work records — what it worked on, the outcome, the evidence class, and whether each record is anchored. History proves what was recorded, never everything that happened.
Usage
treeship history <agent> [--hub <url>] [--class <class>] [--since <ts>] [--limit <n>]| Option | Description |
|---|---|
--hub <URL> | Pull from a Hub instead of the local store |
--class <CLASS> | Filter to self | runtime | countersigned |
--since <TIMESTAMP> | Only records closed at/after this RFC 3339 timestamp |
--limit <N> | Max records, newest first (default 50) |
--format json | Machine-readable output |
A bare name is normalized to agent://<name>.
What each row shows
Artifact ID, headline, outcome, attestation class, close time, action count, tools-exercised count, signature (verified/unverified against your trust roots), and the anchor state (anchored ✓ / anchored ✗ INVALID / blank).
Two honesty rules:
- Remote records are graded, not trusted. Each hub-served envelope is signature-verified locally, and each hub-claimed Merkle anchor is re-proved offline. Typed fields are read from the verified envelope only.
- Class is capped to evidence. A record can't claim
countersignedwithout the countersignature actually present — records with no supportable class display?.
JSON output
{
"agent": "agent://deployer",
"total": 12, "shown": 12, "ok": true,
"records": [
{
"artifact_id": "art_…", "headline": "fix auth bug", "outcome": "completed",
"attestation_class": "runtime", "closed_at": "2026-07-16T17:00:00Z",
"action_count": 28, "tools_exercised_count": 4,
"signature": "verified", "anchored": "anchored ✓"
}
]
}Exit codes
| Code | Meaning |
|---|---|
0 | Clean (unverified signatures are reported, not fatal) |
1 | Hostile: any record whose claimed anchor fails inclusion re-verification (anchored ✗ INVALID) |
See also: treeship profile for the aggregated track record, and GET /v1/agents/history for the Hub side.
audit
Audit an agent's transparency log from a Hub — re-verify every anchored entry's inclusion offline, check completeness against the committed anchor, and detect forks, regressions, and rewrites.
profile / verify-profile
A checkpoint-pinned, recomputable track record over an agent's work history. Every number recomputes from the log — a mismatch is a provable lie.