Skip to content
cccskills
FreePublished on GitHub

operator-approval-loop

Set up human approval for agent-drafted messages to customers or partners. Bind decisions to exact text and destinations, record delivery evidence, and prevent outdated approvals or competing workers from triggering unintended sends.

OriginalShow the original description

Operator approval contract with internal filing notices for agent-drafted outbound messages, hashed drafts, epoch-keyed decisions, durable delivery claims and receipts, and a pre-draft baseline gate. Use when an agent drafts messages to external counterparties and a human operator must approve, reject, or steer each send before it leaves.

See how to install

Good for

  • Require approval before sending customer replies
  • Keep an audit trail of approved text and delivery
  • Prevent old approvals from releasing revised drafts
  • Check existing contract facts before filing a draft

Files included (3)

  • SKILL.md12.8 KB
  • references/approval_claims.py7.5 KB
  • references/approval-ledger.sql14.0 KB

SKILL.md (original)

Review the instructions the agent will receive before you install.

Operator Approval Loop

An agent that talks to external counterparties should never send on its own judgment and should keep the operator informed internally. This skill defines the contract: every outbound draft is filed as an obligation, an operator decides on the exact text, and a delivery ledger proves what went out.

When to Use

  • An agent drafts replies to customers, suppliers, investors, or partners in a shared channel, email, or chat, and a human must approve before send.
  • You need an audit trail that links each sent message to the exact draft text, the operator who approved it, and the decision time.
  • You have seen a stale approval release a rewritten draft, or two workers deliver the same approved message twice.
  • Drafts keep re-asking counterparties for facts the ledger already holds.

How It Works

Objects

ObjectMeaning
ObligationOne thing we owe a counterparty. Status moves drafted, then approved or rejected, then sent. Carries direction, counterparty, channel, and an updated_at epoch.
DraftSidecar row holding the exact draft text, a sha256 of that text, origin coordinates (platform, channel, thread, user), and priority (P0 to P3). One per obligation, replaced on re-file.
DecisionAn operator's approve or reject, recorded with the operator id, a nonce, and the draft epoch it was made against.
Approval snapshotImmutable text, hash, epoch and destination recorded by the already-authorized decision writer. Missing snapshots cannot grant dispatch.
ClaimDurable reservation with a random token and state; at most one active claim per obligation.
DeliveryLedger row proving one send or notice for one (obligation, decision) pair.

The reference schema is in references/approval-ledger.sql.

Filing a draft

  1. Clean inputs. Strip control characters, collapse whitespace in single-line fields, and enforce length caps (draft, summary, context, counterparty). Empty or oversized fields are refused, not truncated silently.
  2. Run the baseline gate (below). It may refuse the filing.
  3. Hash the draft text with sha256. The hash prefix goes into the summary so the approval panel shows which text it is approving.
  4. Upsert. If an open drafted obligation already exists for the same (counterparty, channel), replace the draft sidecar and advance the obligation's updated_at. That advance is the epoch rotation: any decision keyed to the old epoch can no longer release the new text. Otherwise insert a new obligation with status drafted.
  5. Route the filing receipt only to a configured, verified internal ops destination. If the origin is that internal destination, acknowledge there. Never-silent means internal reporting, not an automatic external reply. Keep draft hashes, approval status, operator identity and workflow metadata out of counterparty-visible channels. Unknown or unclassified origins stay quiet; a direct message is not automatically internal.

If a verified internal destination is unavailable, retain the filing result in the internal tool result or operator surface. Never fall back to an external or unknown origin. A tool result exposed to outsiders is not an internal surface.

Filing a draft does not authorize an external response. Any policy-permitted clarifying question or neutral response is a separate outbound decision, subject to the existing mention, channel, draft-only, frozen and never constraints in counterparty-channel-discipline. It must not disclose internal approval metadata.

Baseline gate

Before any draft is filed, query the current baseline for the counterparty (a temporal ledger, contract store, or CRM):

  • Signed or delivered contract on record: refuse the filing with the evidence and a recommendation. Asking a counterparty about specs after signing is the exact failure this gate exists to stop.
  • Operator override: force_despite_signed_contract lets the filing through and stamps [BASELINE_OVERRIDE_SIGNED_CONTRACT] into the draft context.
  • Gate service unreachable: the filing proceeds and the context is stamped [BASELINE_CHECK_UNAVAILABLE]. The panel sees that the guard was off. Failures never silently disable the gate.
  • When facts are available, attach the freshest few to the context as a [BASELINE FACTS: ...] digest so the draft lands with current truth.

Deciding

The approval panel lists obligations with status drafted and direction we_owe_them. Approve or reject writes a decision row carrying the draft epoch (draft_updated_ts) and flips the obligation status in the same transaction. A decision whose epoch does not match the current updated_at is stale and must not release anything.

For an already-authorized approve decision, the same transaction inserts an immutable obligation_approval_snapshots row: decision and obligation IDs, current draft epoch, exact text and SHA-256, platform/channel/thread, and kind draft_sent. The decision writer must establish authorization before writing; the reference never authenticates an operator or manufactures a decision. Automatic approval policy is not enabled or expanded by the reference. Legacy decisions without snapshots require explicit reconciliation or a new approval; never backfill permission from the current mutable draft.

Delivering

The SQLite reference is references/approval_claims.py. It grants dispatch permission but never calls transport. Use an existing local reference database initialized from the SQL fixture; the module does not apply schema or production migrations. Only a trusted decision writer may populate approval records. All writers must enable foreign keys and recursive triggers and honor the schema guards; administrative database tampering is outside this model.

  1. Discover bound approved drafts. Discovery is not permission. claim() opens its own BEGIN IMMEDIATE transaction, validates the current approved epoch, exact text, computed SHA-256 and full destination against the snapshot, and inserts a unique claim before returning its token. A conflict stops the worker before transport. Completed receipts cannot be claimed again.
  2. begin_dispatch() revalidates the binding and atomically changes claimed to dispatching using the token. Only its winning caller receives the exact draft_text and destination after commit. Never regenerate text, reread a mutable sidecar for transport, or reuse the payload for another attempt. A nested caller transaction is refused; permission cannot depend on a later caller commit. No database transaction remains open across transport.
  3. A confirmed successful result goes to complete(), which atomically records the delivery coordinate, marks the claim delivered and flips the obligation to sent. Identical completion is a no-op; conflicting coordinates fail. The receipt UNIQUE key deduplicates records, not prior external effects.
  4. Exceptions, timeouts, worker death after begin-dispatch, or failed receipt persistence leave a blocked attempt. mark_unknown() records uncertainty. Unknown claims never expire, reopen, auto-retry or allow another decision for that obligation to bypass them. A trusted caller may use reconcile() with confirmed successful coordinate and evidence; the module does not verify that evidence. An absent receipt is not proof of non-delivery.

The guarantee is one automatic dispatch attempt per approved decision, not exactly-once external delivery. A crash after begin-dispatch but before transport can leave zero sends and a held claim. Releasing an unknown outcome for a new attempt would require fencing the original executor and verifying provider semantics; this reference deliberately provides no such retry operation.

Claim stateAllowed next states
claimeddispatching or cancelled before dispatch
dispatchingdelivered or unknown
unknowndelivered through trusted reconciliation only
delivered, cancelledterminal; decision key cannot be reused

While a claim is active, database guards freeze obligation, draft and decision writes, including replacements. Snapshots and claims cannot be erased. Cancel a claimed operation with its token before re-filing; the stale token then grants nothing. After dispatch begins, hold new edits or revocation for reconciliation. This serializes changes instead of pretending to recall an in-flight operation.

Rejected decisions and legacy rows without draft sidecars/snapshots never enter this external draft-send path. Report them on the internal operator surface for manual handling. Internal receipt footers remain internal: approved by <operator> · receipt <decision_id> · draft sha256 <prefix>. Never alter already-approved external text to append workflow metadata.

Focused local validation uses temporary databases, separate connections and a simulated attempt counter, not a provider or real message: python3 -m unittest discover -s tests/skills -p 'test_approval_delivery_claims.py'. The tests require Python 3.11+ with SQLite serialization support; the reference uses only the standard library. The existing desk-pattern contract checks remain a separate compatibility check.

Time-boxed auto-approval (optional)

A draft may carry auto_send_after (epoch seconds). A sweep approves drafts whose deadline passed with no decision, recording operator auto-ttl, then delivery proceeds through the normal path. Operator actions always win: a decision flips status before the sweep sees it, and a re-file rotates the epoch and moves or clears the deadline. The sweep re-checks status and epoch inside the write transaction so a race resolves as a no-op. Drafts without a deadline stay hard-gated forever.

Signal linkage

A draft can name the inbound obligation it answers (signal_obligation_id). This is the only truthful link for latency measurement (inbound signal to drafted response) and lets the SLA scan treat that inbound item as answered. Reject the filing if the referenced row does not exist.

Examples

File a draft

file_request(
  draft="Thanks, we can hold the slot until Friday. Which start date works?",
  counterparty="acme-supplier",
  context="reply to delivery window question",
  origin_platform="slack", origin_channel="#acme-shared",
  origin_thread="1712345678.000100", priority="P1",
  signal_obligation_id=412)
-> {obligation_id: 431, draft_sha256: "9f2c...", refiled: false}

The configured, verified internal ops destination sees: Draft filed for approval (P1, sha 9f2c8a1b). Waiting on operator. The counterparty-visible origin channel receives no filing notice. If no verified internal destination is available, the receipt stays in the internal tool result or operator surface, with no external fallback.

Re-file after a steer

The operator asks for a shorter draft. Filing again for the same (counterparty, channel) returns refiled: true, the sidecar text and hash change, and updated_at advances. An approve clicked on the old panel row carries the old epoch and is ignored.

Gate refusal

DeskApprovalError: baseline gate refused this draft: the ledger shows a
signed contract for 'acme-supplier'. Evidence: master agreement executed
2026-08-14. Recommendation: do not ask. Re-file with
force_despite_signed_contract=true if this is genuinely a new thread.

Delivery footer in an internal channel

Confirmed for Friday, start date 2026-09-08.
approved by operator-a · receipt 118 · draft sha256 9f2c8a1b2d3e4f50

Invariants to test

  • Filing receipts go only to configured, verified internal ops; the origin receives one only when it is that verified internal destination.
  • An unknown origin stays quiet. An unavailable internal destination uses the internal tool result or operator surface, with no external fallback.
  • Same (counterparty, channel) filed twice yields one obligation, two epochs.
  • A decision with a stale epoch never results in a delivery row.
  • Two concurrent claimants yield one dispatch permission; losers never attempt transport.
  • Unknown outcomes and failed receipt persistence never enable an automatic retry.
  • Successful completion records the receipt and sent status in one transaction.
  • An altered epoch, text, hash or destination cannot acquire or begin a claim.
  • Active claims block re-file; only pre-dispatch cancellation can release that hold.
  • Gate unavailable stamps the marker; gate signed refuses without force.
  • Auto-ttl never fires against text the operator has since re-filed.

Reviews

No reviews yet. Tell others what you thought.

Comments

No login needed. Share how it worked for you, ask a question, or note where it did not run.

No comments yet. Be the first to write one.

Comments are published right away. Do not include personal information such as email addresses or phone numbers. Up to 2 URLs.

0 / 1,000 characters

This form is protected by reCAPTCHA, and the Google Privacy Policy and Terms of Service apply.

More skills from this repository

accessibility

FreeSummary

Design and audit interfaces for Web, iOS, and Android using WCAG 2.2 accessibility guidance. Check screen-reader labels, keyboard navigation, contrast, and control sizes to identify barriers and guide implementation for people using assistive technology.

  • Adding descriptive labels to icon-only buttons
  • Reviewing keyboard navigation and modal focus
  • Checking contrast and control sizes
affaan-m/ECC283.4KUpdated Oct 16, 2026

Investigate unreliable AI agents across 12 layers, including instructions, memory, tool execution, and display. Use code and logs to trace failures, rank findings by severity, and produce an ordered repair plan that prioritizes code changes.

  • Investigating failures that appear only inside an agent application
  • Tracing old conversation content leaking into new sessions
  • Checking skipped or falsely claimed tool execution
affaan-m/ECC283.4KUpdated Oct 16, 2026

agent-eval

FreeSummaryAI rating 80

Compare coding agents on tasks from your own codebase. Define success checks, run agents in separate Git working directories, and report pass rates, available API costs, completion time, and consistency across repeated trials.

  • Comparing coding agents on real development tasks
  • Evaluating a new tool or model before adoption
  • Checking performance after agent updates
affaan-m/ECC283.4KUpdated Oct 16, 2026

Design and refine an AI agent’s available actions, tool inputs, and result formats. Define recovery steps for errors, manage the information kept in context, and evaluate changes through completion rates, retries, and cost per successful task.

  • Define an AI agent’s tool set and input formats
  • Make tool results and next steps easier to interpret
  • Specify safe retries and stopping conditions
affaan-m/ECC283.4KUpdated Oct 16, 2026

Diagnose AI agent runs that fail repeatedly, loop through tools, or lose track of the task. Capture errors and environment details, test a focused recovery action, and produce a report explaining the evidence and outcome.

  • Diagnosing repeated failures in an agent run
  • Investigating loops through the same tool calls
  • Reviewing output degradation as context grows
affaan-m/ECC283.4KUpdated Oct 5, 2026

agent-introspection-debugging

FreeSummaryAI rating 100

Diagnose failed AI agent runs by capturing errors, checking likely causes, and choosing small, reversible recovery actions. The workflow produces a structured report of the failure, recovery steps, and evidence, helping explain repeated retries or stalled progress.

  • Diagnosing repeated tool calls with no progress
  • Investigating quality loss as context grows
  • Checking file paths and environment mismatches
affaan-m/ECC283.4KUpdated Oct 16, 2026

See all skills by affaan-m

Report a problem with this skill