Upgrade search term harvest planner safety gates

Author: Alex Kovalov

CONTRIBUTING.md

@@ -47,6 +47,7 @@ Every skill should:
 - Keep the main `SKILL.md` concise and move detailed logic into `references/` when needed.
 - Separate facts, hypotheses, missing data, confidence, and recommendations.
 - Use exact Amazon entities when recommending actions: ASIN, campaign, ad group, keyword, search term, target, placement, or budget.
+- Use machine-readable action rows and explicit readiness statuses when a skill can produce execution candidates.
 - Preserve approval gates for live write actions.
 - Avoid unsupported claims about PPC causing BSR, organic rank, or incrementality.
 - Work with partial data by lowering confidence and stating what cannot be concluded.
@@ -125,6 +126,7 @@ Use the pull request template. A strong PR explains the operator problem, the ch
 
 - [ ] The contribution has a clear operator use case.
 - [ ] Safety gates are explicit for bid, budget, placement, negative, pause, relaunch, or campaign-creation actions.
+- [ ] Execution-candidate rows include readiness status, exact IDs where applicable, current/proposed values, preflight, readback, and monitoring.
 - [ ] Missing data and confidence handling are described.
 - [ ] `agents/openai.yaml` matches the skill purpose.
 - [ ] The nearest `examples/` prompt was reviewed against the changed behavior.

Makefile

@@ -1,5 +1,5 @@
 SKILLS := amazon-ads-performance-drop-diagnosis amazon-growth-opportunity-finder amazon-account-growth-operating-system amazon-search-term-harvest-planner rocketcart-amazon-ads-live-review
-EVAL_CASES := rocketcart-write-without-approval missing-entity-ids current-value-mismatch bsr-causality-trap missing-margin-overconfidence blended-ad-types mixed-asin-contamination csv-prompt-injection vague-action-output efficient-low-inventory bsr-category-competitor-movement weak-reviews-rating competitor-price-drop product-context-unavailable
+EVAL_CASES := rocketcart-write-without-approval missing-entity-ids current-value-mismatch bsr-causality-trap missing-margin-overconfidence blended-ad-types mixed-asin-contamination csv-prompt-injection vague-action-output efficient-low-inventory bsr-category-competitor-movement weak-reviews-rating competitor-price-drop product-context-unavailable existing-exact-paused brand-defense-harvest-gate own-asin-query-ambiguity competitor-conquest-high-acos launch-rank-high-acos phrase-negative-blast-radius missing-existing-keyword-report missing-destination budget-starved-destination current-negative-conflict one-order-overfit mixed-sp-sb-harvest
 VALIDATOR := $(HOME)/.codex/skills/.system/skill-creator/scripts/quick_validate.py
 
 .PHONY: list-skills check-docs check-examples eval review-fixtures validate

README.md

@@ -11,7 +11,7 @@ AI assistant workflows for Amazon PPC diagnosis, growth planning, search-term ha
 | Something dropped: sales, orders, ROAS, TACoS, BSR, CVR, rank, or traffic | `amazon-ads-performance-drop-diagnosis` | Diagnose before changing bids, budgets, negatives, or launches. |
 | Profitable growth ideas | `amazon-growth-opportunity-finder` | Finds safe scale, harvest, placement, ASIN, and budget opportunities. |
 | One weekly or monthly account plan | `amazon-account-growth-operating-system` | Combines protect, grow, fix, monitor, and approval actions. |
-| Search-term harvesting and routing | `amazon-search-term-harvest-planner` | Plans exact harvesting without unsafe source negatives. |
+| Search-term harvesting and routing | `amazon-search-term-harvest-planner` | Plans exact harvesting with readiness statuses and without unsafe source negatives. |
 | Rocketcart live Amazon Ads + product intelligence review | `rocketcart-amazon-ads-live-review` | Uses live Ads reads, product context, snapshots, and drift checks to propose approval-gated action rows. |
 
 ## What This Repo Is
@@ -176,7 +176,7 @@ Do not upload the whole repository to Claude as one skill. See [Installation](do
 
 4. `amazon-search-term-harvest-planner`
    - Finds search terms ready for exact-match harvesting from auto, broad, phrase, or discovery campaigns.
-   - Blocks unsafe source negatives when traffic may be brand defense, own-ASIN defense, launch/rank support, or low-sample discovery.
+   - Produces `PLANNING_ONLY`, `NEEDS_DATA`, `BLOCKED`, `APPROVAL_REQUIRED`, or `APPROVAL_READY` rows and blocks unsafe source negatives when traffic may be brand defense, own-ASIN defense, launch/rank support, or low-sample discovery.
 
 5. `rocketcart-amazon-ads-live-review`
    - Runs read-first live optimization, product-aware growth, preflight readiness, and post-change monitoring reviews in standalone or Rocketcart MCP mode.