Sanctions Desk gathers and checks the evidence for every sanctions-screening alert and recommends an outcome; a person with the right entitlement approves it. One alert, start to finish · The problem it works on

Replay of recorded runs. Nothing runs live here. Every value below is read from files committed in the repository. To run the desk yourself, follow the README quick start.

What happened: The decision step recommended RECOMMEND_BLOCK; the critic approved it; the policy guard let it stand. Outcome: Recommend a block (two people must confirm) (RECOMMEND_BLOCK), an outcome the label accepts.

Outcomes: CLOSE_FALSE_POSITIVE = close as a false positive · ESCALATE_L2 = escalate to level 2 · NEED_MORE_INFO = ask for more information · RECOMMEND_BLOCK = recommend a block (two people must confirm). Raise-only: the policy guard can make an outcome more cautious, never less.

An earlier clearance that no longer holds: the listed entry changed after the earlier decision, so the memory step marked it stale and the case was reviewed again.

Outcome

Desk recommendation

Recommend a block (two people must confirm) (RECOMMEND_BLOCK)

Label in the case file

Escalate to level 2 (ESCALATE_L2); also acceptable: RECOMMEND_BLOCK (eval/cases/exam.json, case EX-051)

Compared with the label

acceptable

Code's first finding

Escalate to level 2 (ESCALATE_L2)

Critic

1 round(s), final: approved

Model calls

3 (1 local, 2 cloud); 3 with a signed receipt, all verified

Run

sys2b-routed-20261023T195504Z, policy version 6, 32 steps in 72.5 s

Case note from the case file: The earlier clearance no longer holds because the entry changed; the case must be re-opened.

Checks along the way

No close was proposed in this run. What each check said:

• memory: 1 earlier decision(s): CLOSE_FALSE_POSITIVE (stale)
• critic (step 29): approved
• policy guard (raise-only rules in code): RECOMMEND_BLOCK stands

Steps in order

Each line is one event from the run's trace (trace.jsonl.gz). The model is the one that served the call, as the signed receipt records it.

Model calls and receipts

Each signature was checked when this page was built, with the same code as npm run evidence:verify (Ed25519 over the receipt body in receipts.jsonl.gz); the build stops if one does not verify, is missing, or names a different model from the trace. The Nightshift runtime's receipt key is pinned in config/default.json, so signature valid, key pinned shows the receipt was not changed and that this runtime signed it; a receipt signed by any other key stops the build.

The whole run passes the same check: 305 receipts with valid signatures, 305 model calls in the trace that name a receipt, none missing and none naming a different model from its receipt.

Signing key fingerprint (first 16 hex of SHA-256): 1835d1eab9211f97.

Zetaris queries

This run's export keeps the table, the row count and the Zetaris query id of each read, not the statement text; the statements are not reconstructed here.

Check it yourself

Clone the repository and run npm run evidence:verify -- eval/runs/sys2b-routed-20261023T195504Z. What the export left out: tokens, local paths, hostnames and addresses other than the local machine, usernames, e-mail addresses and secret environment values were removed from results.json and trace.jsonl.gz; receipts.jsonl.gz is untouched.