Menú
Signed session receipts for every OpenClaw task (v0.9.4)
Instalar con Codex o Claude Copia este prompt, pégalo en Codex, Claude u otro asistente, y deja que revise la página de la skill y la instale por ti.
Basado en la clasificación ocupacional SOC
| name | treeship |
| description | Signed session receipts for every OpenClaw task (v0.9.4) |
| version | 0.9.4 |
| author | Zerker Labs |
| homepage | https://treeship.dev |
| openclaw | {"category":"security","tags":["attestation","receipts","verification","trust","mcp"]} |
| requires | {"bins":["treeship"],"env":["TREESHIP_ACTOR"]} |
Generate signed, verifiable session receipts for every OpenClaw task. Each receipt captures what happened, who did it, what changed, and provides cryptographic proof that the record is trustworthy.
Current version: 0.9.4 — Rust core, 161 tests passing, Ed25519 DSSE envelopes, offline verification.
One-liner (macOS/Linux):
curl -fsSL treeship.dev/setup | sh
Or via npm (inspectable, signed package):
npm install -g treeship
treeship init
The setup script auto-detects OpenClaw and installs this skill to ~/.openclaw/skills/treeship/.
treeship session start --name "<task description>" --actor agent://openclaw
Optional: set model metadata for richer receipts:
export TREESHIP_ACTOR="agent://openclaw"
export TREESHIP_MODEL="<your model>"
Every shell command MUST be prefixed with treeship wrap --:
treeship wrap -- npm test
treeship wrap -- cargo build --release
treeship wrap -- python3 script.py
treeship wrap -- curl -sS https://api.example.com/endpoint
For multi-step shell commands:
treeship wrap -- bash -c "git add -A && git commit -m 'fix: resolve bug'"
For sub-tasks within a larger task, continue using the same session. Do NOT start a new session. The receipt should capture the entire task as one unit.
Close with a meaningful summary:
treeship session close \
--headline "<one-line accomplishment>" \
--summary "<detailed description of what changed>" \
--review "<what a reviewer should check before trusting this work>"
Upload for a shareable URL:
treeship session report
Good headline: "Fixed JWT expiry bug that caused 401s after 24 hours" Bad headline: "Fixed a bug"
Good summary: "Found the TTL was set to 24h instead of 7d in token.ts. Fixed the constant, added a regression test, and confirmed the fix in local tests." Bad summary: "Fixed the bug."
Good review: "Verify the TTL change works with existing tokens in production." Bad review: "Please review."
| Data | How | Status |
|---|---|---|
| Shell commands | treeship wrap -- | AUTO |
| Exit codes | treeship wrap -- | AUTO |
| Elapsed time | treeship wrap -- | AUTO |
| File writes | File snapshot diff | AUTO |
| Model name | TREESHIP_MODEL env var | EXPLICIT |
| Actor URI | TREESHIP_ACTOR env var | EXPLICIT |
| Session narrative | --headline/--summary/--review | EXPLICIT |
| Merkle proof | Automatic | AUTO |
| Ed25519 signatures | Automatic | AUTO |
treeship wrap --treeship daemon start for atime detection)If @treeship/mcp is configured as your MCP server, every callTool invocation is automatically attested:
No treeship wrap -- needed for MCP tool calls. Configure via:
npx -y @treeship/mcp@latest setup
treeship status # show keys, recent artifacts, hub status
treeship verify last # verify the most recent artifact
treeship verify <id> # verify a specific artifact
treeship package verify <path> # verify a .treeship package offline
treeship hub attach # connect to treeship.dev Hub
treeship hub push <id> # push artifact to Hub
treeship ui # interactive TUI dashboard
treeship doctor # diagnostic check
treeship add # instrument AI agents in current project
treeship quickstart # guided first-time setup
Verify any receipt locally (pure WASM, no network):
treeship verify art_<id>
treeship package verify .treeship/sessions/ssn_*.treeship
Or via web: https://treeship.dev/verify/
treeship --help, treeship <command> --helpAttestation skill for lobster.cash wallet operations
Use when the user asks for "the URL", "a shareable link", "the proof I can send to <someone>", or otherwise wants the current Treeship session published as a session report on treeship.dev.
Create cryptographic attestations for AI agent actions using Treeship.dev. Use when the user wants to sign agent actions, create verifiable receipts, wrap commands with attestations, verify artifact chains, push artifacts to the Treeship Hub, integrate Treeship into their agent workflow, or capture model/token provenance. Covers the CLI (treeship), Python SDK (treeship-sdk), TypeScript SDK (@treeship/sdk), MCP Bridge (@treeship/mcp), and Hub API (api.treeship.dev). Trigger on any mention of attestation, verifiable receipts, signed receipts, agent trust, Treeship, action signing, cryptographic proof of agent work, model provenance, or token attestation.
Use when operating Hermes Agent in a repo or organization where AI work should produce Treeship-verifiable receipts. Captures Hermes sessions, MCP-routed tool calls, wrapped shell commands, approvals, handoffs, and publishable session reports while avoiding raw secrets in receipts.
Use Treeship (treeship.dev) from Perplexity Computer — install, configure, sign agent actions, verify receipts, push to the Hub, run approval-gated workflows, and work on the zerkerlabs/treeship source repo via GitHub. Trigger on any mention of Treeship, treeship.dev, signed receipts, agent attestation, verifiable agent actions, cryptographic proof of agent work, DSSE artifacts, Merkle proofs, approval-gated actions, session receipts, or the treeship CLI. Also load when asked to contribute to or inspect the zerkerlabs/treeship GitHub repository.
Create cryptographically signed, portable trust receipts for AI agent workflows using Treeship.dev. Use when the user wants to sign agent actions, create verifiable receipts, wrap commands with attestations, verify artifact chains, push artifacts to the Treeship Hub, manage approval-gated actions, or integrate Treeship into their agent workflow. Covers the Treeship CLI (treeship), Python SDK (treeship-sdk), TypeScript SDK (@treeship/sdk), MCP bridge (@treeship/mcp), and Hub API (api.treeship.dev). Trigger on any mention of attestation, verifiable receipts, signed receipts, agent trust, Treeship, action signing, cryptographic proof of agent work, portable receipts, approval workflows, session receipts, Merkle proofs, decision cards, or coverage levels.