- Home
- Skills
- Trading & Investing
- scarbook decision auditor
Works with the AI Agents you already use
scarbook decision auditor
Audit Scarbook decisions against verified evidence provenance and system constraints for deterministic action-gating.
$29
scarbook decision auditor
Example session with this skill installed
Audit this decision for BTC/USD. The payload includes 3 evidence receipts and a 'no_new_position' constraint. Check if the hash at receipt_id 'rx_99' matches the source data.
- Read your context and instructions
- Compiled the scarbook decision auditor
I have completed the audit for the BTC/USD decision payload. The audit identified a critical provenance failure regarding the evidence hashes provided in the decision object.
Audit Result: FAIL
| Check Category | Status | Details |
|---|---|---|
| Overall Status | fail | Audit rejected due to hash mismatch and constraint violation. |
| Evidence Provenance | fail | receipt_id: rx_99 hash mismatch. Computed: e3b0c4..., Provided: a1b2c3... |
| Evidence Trust | pass | All 3 observations marked as status: ok. |
| Constraint Check | fail | Violation: no_new_position is active. Action OPEN_LONG rejected. |
| Symbol Alignment | pass | All evidence links match BTC/USD. |
| Logic Consistency | review | Missing RSI value in rx_102; preserved as null. |
Audit Findings
The decision payload failed validation on two counts. First, the hash for evidence rx_99 does not match the canonical fields, indicating the data may have been altered or corrupted since the observation. Second, the proposed action conflicts with the high-priority no_new_position constraint currently active in your configuration. The audit has preserved the missing RSI value from rx_102 rather than defaulting, as per the operating contract.
Next steps
- Recalculate the evidence hash for
rx_99to ensure data integrity. - Check the active constraint set to determine why
no_new_positionis currently enforced. - Resubmit the decision payload once the action is aligned with system constraints.
Connects securely to your tools. The creator never sees your data.
About this skill
Manual trade execution often lacks the rigorous verification needed to prevent errors in high-stakes environments. This skill audits Scarbook decision payloads to ensure every proposed action is backed by verified evidence and adheres to active system constraints. It provides a deterministic gatekeeper for provenance and trust without ever mutating state.
What it does
- Verify provenance by recomputing hashes for every linked evidence observation to detect tampering.
- Validate evidence trust by checking observation statuses and rejecting any data marked as unavailable or erroneous.
- Enforce constraints using Scarbook priority logic to reject actions violating position limits or blocklists.
- Audit logic to ensure decisions, symbols, and timestamps align across the entire payload.
- Detect mixed-mode risks by flagging payloads that attempt to combine live data with fixture or malformed observations.
How it works
- Load payload containing the decision object, evidence array, and current constraints.
- Execute audit via the internal scripts to validate hashes and deterministic rules.
- Review results based on the status envelope: pass, review, or fail.
Frameworks & tools
The skill utilizes a Python-based implementation (scripts/audit_decision.py) and adheres to the machine-readable definitions in references/contract.md.
Why this beats prompting it yourself
LLMs struggle with deterministic hash verification and strict boolean logic over large JSON payloads. This skill offloads the heavy lifting to validated scripts, ensuring zero-tolerance for evidence tampering or constraint violations.
Use cases
- Gatekeeping actions by verifying a decision before passing it to an execution engine.
- Post-trade forensic auditing to confirm that past decisions were backed by valid evidence at the time of entry.
- Provenance checks to ensure data integrity across distributed trading systems.
Known limitations
This is a read-only audit tool. It cannot place trades, modify existing receipts, deactivate system constraints, or update decision state.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Trading & Investing.
Read-only tool. Does not support trade execution or state mutation. Requires input to follow the Scarbook JSON contract for decisions and evidence.
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 buying a skill.
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 21 days ago
- Software
Python