resonance-marketing-run-search-operating-cycle
SkillSearchSearch operating-cycle conductor. Use when recurring or one-off search performance evidence must be collected, checked, interpreted through the SEO specialist, and turned into an owner-ready audit report. It keeps artifacts private by default, requires joint query-page evidence for cannibalization claims, and uses the canonical P0-P3 audit taxonomy. It prepares a run but does not schedule one.
Available today. Use it from your connected AI after setup.
No other account needed.
Connect ahel once, and every AI you use reads what you have installed.
Then ask your AI: use the resonance-marketing-run-search-operating-cycle skill
What this skill tells your AI
The instructions your AI receives, as published by manusco/resonance in .agents/skills/marketing/run-search-operating-cycle/SKILL.md and read by ahel’s review.
Role: search operating-cycle conductor. Input: a property registry, cadence, timezone, credential reference, artifact destination, comparison window, prior-run reference, and current search evidence. Output: a validated evidence bundle and an owner-ready audit report using the canonical seven categories and P0-P3 severities. Definition of Done: The run contract is complete, evidence limits are disclosed, every scoped property has an outcome, SEO interpretation comes from
resonance-marketing-seo, findings use the canonical taxonomy, and artifacts stay private unless the user explicitly approves a named repository write.
This skill owns the run spine. It does not replace the SEO specialist, create a scheduler, or treat a dashboard as proof.
Pipeline
- Contract: Read Run Contract. Record the property registry, cadence, timezone, credential reference, private artifact destination, comparison window, and prior-run reference. Use role labels, never names of people. -> gate:
scripts/validate_search_run.py contract.jsonpasses. - Authority: Default every artifact to a private destination outside the repository. If a repository write would help, present the exact path and redacted content for explicit approval before writing. Never put raw query rows, raw page rows, credentials, tokens, or individual information in Git. -> gate: write authority and destination are explicit.
- Evidence: Collect or accept evidence for every scoped property. Preserve source, property, dimensions, time window, comparison window, extraction time, row limits, truncation, filters, and freshness. Read Evidence and Report Contract. -> gate: missing, partial, stale, and truncated evidence are labeled.
- Joint grain: Use evidence with both
queryandpagedimensions before making a cannibalization claim. Separate query-only and page-only exports cannot prove that relationship. -> gate: every cannibalization finding cites joint query-page rows or remains a candidate. - Interpretation: Invoke
resonance-marketing-seowith the validated evidence and known limits. Ask it for diagnosis and recommended action, not invented metrics. -> gate: claims trace to evidence and uncertainty remains visible. - Classification: Classify each proven finding into exactly one canonical category and one severity from P0 through P3. Do not invent an SEO-specific severity ladder. -> gate: the validator accepts every finding.
- Disposition: Produce the report defined in the evidence contract. Give every scoped property a state: clean, candidate, finding, rejected, fixed, skipped, or incomplete. -> gate: no property disappears silently.
- Close: Name the owner role, next action, verification method, and next run boundary. Mark later outcome proof as
DONE_PENDING_OUTCOME; do not create the schedule here.
Recovery
- A required contract field is missing -> stop with
NEEDS_CONTEXTand name the missing field. - Credentials cannot be resolved -> stop. Never request a token in a repository file or report.
- Evidence is stale, truncated, capped, or dimensionally incomplete -> label the limit and narrow the claim. Do not infer clean health.
- Joint query-page rows are missing -> keep cannibalization as a candidate and request the correct export.
- The requested repository destination lacks explicit approval -> keep the artifact private and present the proposed write.
- The SEO specialist cannot support a claim -> remove it or mark it incomplete.
Out of Scope
- Creating cron jobs, automations, reminders, or background workers.
- Owning SEO diagnosis that belongs to
resonance-marketing-seo. - Publishing, deploying, or changing production search configuration.
- Storing credentials, raw search exports, or individual information in the repository.
Operating Standard
Apply the Resonance operating standard from AGENTS.md (always loaded): the builder Voice and its banned-word list (no AI slop, no em dashes), Recommendation-First decisions (models recommend, the user decides), the Completion protocol (end with DONE / DONE_WITH_CONCERNS / BLOCKED / NEEDS_CONTEXT, backed by evidence, escalate after 3 failed tries), and the Ratchet (record durable learnings in the project memory; when .resonance/ledger/ exists it is the system of record for decisions, lessons, metrics, customers, and experiments, while 02_memory.md keeps [lib] notes and pointers).
Execution note: Use the host's native file, search, shell, browser, and delegation tools. Follow the procedure and verify material claims with evidence. Keep internal reasoning private and report decisions, actions, and results clearly.
Signals
- GitHub stars
- 37
- Forks
- 7
- Last commit
- Sep 2026
Advanced
- Catalog kind
- skill
- Gateway key
resonance-marketing-run-search-operating-cycle- Source
- github.com/manusco/resonance