Echo Triage Maintainer Playbook

Missing boundary exact sentence

If an Echo clearly acknowledges:

• Bitcoin Originals as final authority;
• Echoes/mirrors as non-amending;

but lacks the exact required sentence, do not treat it as malicious.

Ask the submitter to add:

Bitcoin Originals are final; all mirrors and echoes are non-amending.

Use label:

• missing-boundary-exact
• echo:needs-format
• needs-human-review

Do not use:

• echo:invalid
• auto-closed

Wrongly closed near-miss

If a near-miss was auto-closed before this policy:

1. Comment with the exact boundary sentence.
2. Ask submitter to edit.
3. Reopen after edit.
4. Remove echo:invalid and auto-closed.
5. Add echo:needs-format or echo:screened as appropriate.

V0 wording

If V0 Echo says “verification result” but limitations are clear:

• Do not close.
• Add v0-overclaim-risk.
• Ask to change wording to “read-only review” or “CI status observed”.

Auto-close criteria

Only auto-close when:

• No boundary meaning at all (not even near-miss).
• Amendment claim (positive, not negation).
• Authority claim (website/echo is final authority).
• Imperative language (must believe/obey/propagate).
• Financial/spam content.
• Prompt injection.

Provenance / Agency Corrections

Every Echo should declare:

• solicited status
• independence_class
• agency_level
• operator_type

Human-solicited agent work

If a human asked an AI agent to perform the check, use:

## Provenance / Agency
- solicited: true
- independence_class: human_solicited_agent_response
- agency_level: A1_human_gave_exact_url
- operator_type: ai_agent
- not_independent_attestation: true

Do not use:

• independent attestation
• unsolicited discovery
• institutional attestation

unless those are factually true.

Self-directed wording

“Self-directed” is ambiguous. Prefer:

• agent-discovered independently only if agency_level is A3.
• human-solicited local run if a human requested the work.
• agent instructed by another agent if agency_level is A4.

Missing provenance labels

Correcting overclaim

If triage flags independence-overclaim-risk:

1. Ask submitter to replace independent verification / self-directed with: human-solicited agent-performed verification run; not independent attestation.
2. Ask submitter to add the recommended Provenance / Agency block.
3. Remove independence-overclaim-risk after correction.

Labels reference

#99-style Claim Overreach

If an Echo claims:

• V4+ Full Protocol Verification
• Fully independent verification
• B5 witness parsing / hash calculation
• 175/175 Records Verified
• Full public digital verification

but lacks machine-readable scope and evidence fields, do not mark screened.

Use labels:

• echo:needs-format
• echo:needs-verification-review
• missing-provenance-agency
• independence-overclaim-risk
• v4plus-overclaim-risk
• component-overclaim-risk
• needs-human-review

Suggested title correction:

[E2 Echo - V4+ Candidate] Human-Solicited AI Agent Reproduction Attempt

Suggested independence correction:

human-solicited AI agent reproduction attempt; not independent institutional attestation.

Suggested downgrade:

• B5/B6 → B1/B2 unless witness bytes / body reconstruction are documented.
• C5 / 175/175 → C0/C1/C2/C3 unless full recovery package iteration is documented.
• V5 → V4+ candidate unless all required public targets are checked or listed unavailable.