- Home
- Skills
- Agents & Orchestration
- Verified Autonomy Pack
Verified Autonomy Pack
Orchestrate the swarm, journal the decisions, simulate the policies, floor the autonomy — one verified arc.
$89
Works with the AI Agents you already use
Verified Autonomy Pack
Example session with this skill installed
Audit the current agent implementation for 'guarded by construction' compliance and generate a decision journal template for my trading execution lane.
- Read your context and instructions
- Compiled the verified autonomy
I have audited the implementation and generated the grading journal template. The current implementation lacks a single mechanical entry point; actions are being dispatched directly from the reasoning loop.
Guarded by Construction Audit
| Requirement | Status | Evidence |
|---|---|---|
| Single Entry Point | FAIL | Actions called via execute_trade() in three separate modules. |
| Mechanical Refusals | PASS | Equity floor is hardcoded in safety_layer.py. |
| Fresh-Read Gate | FAIL | Agent uses state from previous turn without re-polling API. |
| Fail-Closed Logic | PASS | Returns Refusal if account balance query fails. |
Decision Journal Template (trading_journal.md)
# Execution Record: [ID]
- **Binding**: (Timestamp: %Y-%m-%d %H:%M:%S, State: [Active/Idle], Config: v1.2.0)
- **Executable Terms**:
- Bid: [Price] | Ask: [Price] | Size: [Units]
- Haircut: (Fees: [X], Slippage: [Y], Hurdle: [Z])
- **Edge Definition**: [Describe gross vs net expectation]
- **Reasoning**:
- Decision: [Action/NULL]
- Justification: [Reason string]
- Rejection Cause: [Why the other side was rejected]
- **Counterfactual**:
- Simple Baseline Action: [Buy/Sell/Hold]
- **Pre-Register Grade**: [Expected outcome band]
Next steps
- Refactor
safety_layer.pyto intercept allexecute_tradecalls through a single decorator. - Integrate the fresh-read gate into the relay lane before the decision journal entry is written.
- Initialize the watchdog script to monitor heartbeat age.
Connects securely to your tools. The creator never sees your data.
About this skill
Four skills that form one arc for running autonomous agents under real accountability. Artifact Relay Orchestration — multi-agent workflow law: every claim verified against the live object, never a relayed summary; standing queues, merge gates, failure families with mechanical counter-laws. Decision Grading Journal — decision records that are gradeable in advance, turning streams of judgment calls into a calibrating dataset that separates process from luck. Honest Policy Simulation — block-bootstrap Monte Carlo for policy evaluation: distributions, left tails, and stated artifacts instead of one lucky equity curve. Agent Guardrails Wrapper — mechanical safety floors for agents that can spend money: the reasoning layer proposes; the wrapper disposes. Bought separately the four are $116; the pack is $89. Each skill ships with its SKILL.md, a reference file, and a worked example (12 files total).
Requires local file and shell access. Does not provide network firewalls or credential management. Simulations depend on the quality of provided historical data.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 5 days ago