BD decision-gate
Tamper-evident decision logging for AI agents, with the one thing a local log can't give you: independent third-party verification. Use when an agent is about to do something it can't undo - send money, release data, deploy config, sign a transaction. Commits a hash-chained record of the decision BEFORE the action fires, so it can't be backfilled to look deliberate. Unlike a self-authored audit log (which is still your own word), this pairs with decision-gate-verifier: an external party that confirms the action matched the claim and signs a receipt anchored on Base that ANYONE can re-derive - a mismatch is a fraud proof, not a complaint. Stdlib-only Python, no dependencies, no server, no telemetry, free forever. TRIGGERS: audit trail, decision log, agent accountability, prove what my agent did, irreversible action, pre-commitment, tamper-evident log, compliance record, third-party verification, why did my agent do that.
Tamper-evident decision logging for AI agents, with the one thing a local log can't give you: independent third-party verification.
As a process D 38/100 · Unfinished process — weak spots: result and completion, when it triggers, inputs and preconditions
How to improve
- Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
- A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.
Guard findings · 0
✓ No critical or high findings
Files scanned: 3. Evidence is masked. Grey chips explain why severity was lowered.
Against the Agent Skills spec
- warning
frontmatter-yamlSKILL.md: the frontmatter is not valid YAML (YAML parse error: Nested mappings are not allowed in compact mappings at line 2, column 14: description: Tamper-evident decision logging for AI agents, with the one thing … ^ ); fields were read line by line. The usual cause is a colon inside an unquoted value
Process rating: all ten parameters 38/100
- 0Result and completion. Does not say what the result is
- 0Inputs and preconditions. Does not say what the process needs to start
- 0Failures and branches. Linear process with no failure handling
- 20When it triggers. No condition that starts the skill
- 30Running it twice. 15 mutating operations with no state check
- 50Steps. 2 steps
- 60Tools and files. Uses tools (bash, python) that frontmatter does not declare
- 100Consistency. Name and required fields are in place
- 100Execution cost. Instruction body is 3589 tokens
- 100Progress reporting. Reports progress
- low 10 top-level sections: this looks like several domains in one skill
Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.
Quality signals
- +5Description has no quoted example phrases that should trigger the skill
- +4Description does not say when NOT to use the skill (false activations)
- +3Description length 933: 120–800 characters recommended
- +3No numbered steps or checklist
- +3Output format is not stated: the model decides each time
- +1No license
- +2Single-language instructions
- +4Structure: 11 headings
- +4Has examples (2 code blocks)
Quality base 70; lint remarks subtract, signals add up to 100. Result: 68.