bitsandtea/postking-skills / skills/postking-seo
postking-seo
VerifiedRun SEO / GEO from seed keywords to published, AI-citation-friendly articles using PostKing — seed, expand, cluster, roadmap, brief, write, and publish.
SEO / GEO
An agentic pipeline that takes a brand from 3–10 seed keywords all the way to published, SEO / GEO-targeted blog articles using PostKing.
What is GEO?
GEO stands for Generative Engine Optimization — writing and structuring content so that AI assistants (Claude, ChatGPT, Perplexity, Gemini, Google AI Overviews) quote and cite it when answering user questions.
Classic SEO optimizes for click-through from a search-results page; success is measured as rank position and CTR. GEO optimizes for being included as a cited source inside an AI-generated answer; success is measured as citation rate in AI responses. The end user may never click — being the quoted authority is the win. Both still matter: SEO drives discovery surfaces that index your content, and GEO determines whether that content survives extraction into an AI answer.
GEO-friendly content follows formatting rules that LLM extractors prefer:
- Definition-first paragraphs — the first 1–2 sentences under each H2 answer the heading as a standalone, quotable claim.
- Headings as questions or definitive statements, not clever wordplay, so hierarchy maps cleanly to extractable Q&A.
- FAQ blocks with explicit Q→A structure (
<dt>/<dd>or<h3>+ paragraph). Highest-yield GEO format. - Source-attributable claims — concrete numbers, named studies, dated stats, primary sources cited inline.
- Schema markup (Article, FAQPage, HowTo) as JSON-LD so crawlers parse intent unambiguously.
- Self-contained sections — each H2 makes sense without scrolling up; LLMs extract sections out of context.
- Comparison tables and lists — easy for LLMs to lift verbatim.
- Authority signals in-prose — founding dates, sample sizes, first-person expert framing.
PostKing's SEO / GEO pipeline produces content optimized for both — keyword-target SEO with GEO-friendly structure (definition-first paragraphs, FAQ blocks, source-attributable claims) baked in by default.
When to Use
Use this skill when a user wants to grow organic/AI-citation traffic from keywords: seeding a topic list, clustering keywords into topical hubs, generating a content roadmap, drafting briefs, writing SEO/GEO-structured articles, or publishing/scheduling them. Also use it for post-publish audit work (gap analysis, competitor content diff) and for generating cluster-linked side pages or standalone comparison pages.
When NOT to Use
Not for one-off social posts, blogs without a keyword/cluster angle, or general brand content — route those to the postking-social/postking skill instead. Not for discovering or registering competitor domains — that's the postking-competitor-intel skill (though its output feeds this one). Not for full multi-channel campaign planning — that's postking-storylines.
Minimal tool subset
This skill needs 15 tools out of PostKing's 200+:
seo_add_seeds— register seed keywords.seo_generate_keywords— expand seeds into a full keyword set (async).seo_categorize— tag keywords by search intent (per-keywordupdatesarray, built fromseo_list_keywordsIDs).seo_generate_clusters— group keywords into topic-pillar clusters (async).seo_list_clusters— list clusters.seo_bulk_approve_clusters— approve clusters to unlock brief generation.seo_generate_roadmap— build prioritized article topics from an approved cluster.seo_list_briefs— list generated briefs.seo_get_brief— inspect one brief.seo_edit_brief— structured edit of a brief (fullbriefDatareplacement and/or status flip).seo_approve_briefs— approve briefs, unlocking article writing.seo_write_article— draft an article from an approved brief.seo_list_results— list generated articles.seo_publish_article— publish or schedule.get_job— poll any async operation.
Granular & optional operations — use these to drive a single step in isolation (e.g. you already have a keyword list and want to skip to clustering) or for post-publish/competitive work, rather than running the full pipeline:
seo_list_keywords— list the brand's scored keyword set (keyword, volume, KD, intent).seo_get_market_settings/seo_set_market_settings— read/configure the (location, language) SEO markets keyword research draws from; prerequisite for a language-scopedseo_generate_keywordscall (see step 2).seo_estimate_research_cost— read-only cost +pairCountpreview for one language before running a language-scoped pull.seo_gap— gap analysis: high-value topics not yet covered.seo_competitor— competitor content diff against a domain.seo_auto_assign_cta— batch-assign a published side-page CTA to blogs.seo_generate_side_page— generate a side page linked to an SEO cluster for topical authority.create_comparison_page— generate a standalone comparison/alternatives page.
Prerequisites
- Authenticated with credits — see the
postking-getting-startedskill. - A brand already onboarded — see the
postking-getting-startedskill. - Optional: competitors registered for competitor-aware clustering — see the
postking-competitor-intelskill.
Procedure
"Run the SEO / GEO flow for my brand. Seeds: 'AI content generation', 'social scheduling', 'brand voice'."
Async tools return { operationId, status } immediately — poll get_job({ operationId }) until state is completed, failed, partially_failed, or cancelled.
-
Seed.
seo_add_seeds({ brandId, seeds: ["AI content generation", "social scheduling", "brand voice"] }). -
Expand.
seo_generate_keywords({ brandId })→ async, pollget_job. Expands seeds into ~100 scored keywords against the brand's default market (Global = a single US/en pull) unless markets are configured otherwise.Multi-market / language-scoped research (optional). Only needed to scope a pull to one specific language, or once the brand's SEO markets already span more than one language (a Pro/Enterprise-tier capability). Flow:
seo_get_market_settings({ brandId })to see what's configured →seo_set_market_settings({ brandId, markets })to set 1–3 curated(locationCode, languageCode)pairs (curated set today: Germany 2276/"de", France 2250/"fr", Spain 2724/"es", Brazil 2076/"pt", US 2840/"en"; or passmarkets: "global"to revert to the US/en default) →seo_estimate_research_cost({ brandId, language })to seepairCountandestimatedCostUsdfor that language →seo_generate_keywords({ brandId, language })scoped to it.seo_get_market_settings/seo_set_market_settingsrequire a paid plan — a free/trial key gets a payment-required error. See Pitfalls for what happens if this ordering is skipped. -
Categorize. There's no brand-wide categorize call — build a per-keyword
updatesarray first. Callseo_list_keywords({ brandId })to get keyword IDs, thenseo_categorize({ brandId, updates: [{ keywordId, intent, userTags? }, ...] })—updatesis required (min 1, max 500 entries), each entry is{keywordId, intent?, userTags?}withintent∈ informational | commercial | navigational | transactional. Informational-intent keywords are prime GEO targets — they're the questions users ask AI assistants directly. -
Cluster.
seo_generate_clusters({ brandId })→ async, pollget_job. Groups keywords into topic-pillar clusters. Cluster pillars become topic-authority hubs that AI engines cite back to. If competitors are registered on the brand, this step automatically classifies each cluster'sproductFitand flags competitor-dominated clusters.Competitor-aware clustering. If the brand has competitors registered first (via the
postking-competitor-intelskill),seo_generate_clustersautomatically injects a "Registered competitors" block into the cluster-labeling step — no separate tool call, it's automatic once competitors exist. For each cluster the labeler then returns:clusterMeta.productFit:"core","adjacent", or"out_of_scope".clusterMeta.archetype: "competitor"(only when the cluster's keywords are dominated by a known rival brand), plusdetectedCompetitorand a pre-built "conquest" brief plan (review article + head-to-head comparison + alternatives listicle, optionally a pricing blog) instead of the generic hub-and-spoke plan.- Inspect these via
seo_list_clusters({ brandId, detail: "medium" })or"full"—clusterMeta.productFit/archetypeonly surface atmedium/fulldetail, notshort. - When you
seo_bulk_approve_clustersaarchetype: "competitor"cluster, it produces the conquest-plan briefs rather than generic ones. - Register competitors first with the
postking-competitor-intelskill.
Manual cluster (optional). To hand-create a named cluster instead of waiting on the auto-clustering pipeline, call
seo_create_cluster({ brandId, name, pillarKeyword?, description? })— no LLM call, starts empty, and is created alreadyapproved(skip step 6 for it). Names must be unique per brand; a duplicate returnscluster_name_taken. Add its first brief withseo_create_custom_brief(see the one-off-brief callout under step 6) passing this cluster'sidasclusterId. -
Review clusters.
seo_list_clusters({ brandId, detail: "medium" })— surface clusters with pillar keyword, keyword count, and top keywords (use"full"for raw data;clusterMeta.productFit/archetypeonly surface atmedium/full). Ask the user which to target, or pick by size/volume. -
Approve.
seo_bulk_approve_clusters({ clusterIds: [...] })(orseo_approve_clusterfor a single one; reverse withseo_reject_cluster/seo_bulk_reject_clusters). Approval is the gate — only approved clusters get briefs drafted, and this fires an async brief-generation job per cluster.One-off brief (optional, skips roadmap). To draft a single custom brief without running
seo_generate_roadmapfirst — e.g. a specific pillar keyword the user names outright — callseo_create_custom_brief({ brandId, pillarKeyword, type, supportingKeywords?, intent?, research?, selectedCompetitors?, clusterId? })(type∈ blog | comparison | tool | landing; async, pollget_job). Pass an existingclusterId(fromseo_create_clusterorseo_list_clusters) to file it under a named cluster, or omit it to auto-mint a throwaway single-brief cluster (origin: "manual"). The resulting brief lands inpending_reviewjust like a roadmap-generated one — continue at step 8. -
Roadmap.
seo_generate_roadmap({ clusterId })— produces prioritized article topics from the approved cluster. Each roadmap item gets a brief auto-drafted in the background, specifying FAQ blocks and definition-first headings to drive the GEO format. -
Review & edit briefs.
seo_list_briefs({ brandId, status: "pending_review" })→seo_get_brief({ briefId })to inspect. Refine withseo_edit_brief({ briefId, briefData })— pull the currentbriefDatafromseo_get_brief, mutate the outline (H2s, FAQs, keyword targets, etc.) locally, and send the full replacement back; there is no free-textinstructionsparam — or rebuild from scratch withseo_regenerate_brief({ briefId })(async). Brief approval is where you lock in the FAQ blocks and definition-first H2 structure — the article writer treats the approvedbriefDataas a contract. -
Approve briefs.
seo_approve_briefs({ briefIds: [...] })— unlocks article writing. -
Write.
seo_write_article({ roadmapItemId })— drafts an article from a roadmap item whose brief is approved. Repeat per item. Articles ship with FAQ schema and definition-first H2 paragraphs by default. -
Review results.
seo_list_results({ brandId })— list generated articles and their status. -
Publish.
seo_publish_article({ articleId, publicationId, schedule })— publish, or schedule for later. This is the free-tier choke point. -
Audit (optional).
seo_gap({ brandId })— uncovered high-value topics.seo_competitor({ brandId, domain })— competitor content diff. -
Auto-assign CTAs (optional, post-publish).
seo_auto_assign_cta({ blogIds: "all" })(or a 1–50 array) — batch-assigns a published side-page CTA to each blog. Skips blogs that already have a CTA and Webflow-synced blogs by default; pass overwrite/include-webflow flags to change that. Hard cap of 50 blogs per call. -
Cluster-linked side pages / comparison pages (optional).
seo_generate_side_page({ slug, key, clusterId })—slugis the parent landing page's slug; generates a side page under it, linked to an SEO cluster, strengthening the cluster's topical authority: it expands covered subtopics, feeding the citation patterns AI engines use when answering questions in that domain.create_comparison_page({ ... })generates a standalone comparison/alternatives page. Auto-assigned CTAs increase conversion yield without changing the article's GEO-friendly structure.
Roadmap from real search data (Search Console / Bing connected)
Once a brand has search data synced, PostKing reconciles the roadmap against the queries people actually used (last 90 days). It runs on first sync and weekly, automatically. It only touches suggested items: approved, writing, published and completed items are never re-ranked.
What it does: re-ranks suggestions by real demand (each carries metadata.searchEvidence: impressions, position, queries); hides items an existing page already covers or that rank top 3 (status ignored, metadata.hiddenReason = already_published / ranks_top3 / merge_loser); imports observed keywords and topics; and adds refresh items.
seo_search_data_status({ brandId? })— connected?, last reconcile (statusgenerating / completed / failed / stale,summary: movedUp / nowCovered / newKeywords / newTopics / refreshItems), and roadmap counts (refreshOpen, hidden, withSearchData). Call first; ifconnectedis false there is nothing to reconcile.stale= a crashed run; start a new one.seo_search_data_reconcile({ brandId? })— run it now instead of waiting for the weekly run. Returns immediately; pollseo_search_data_statusuntilreconcile.statusiscompleted. 400 = no data synced yet, 409 = already running.seo_list_roadmap({ type: "refresh" })— pages to rewrite. A refresh item points at an existing page (metadata.refresh.pageUrl) and carries its weak queries plus the AI-assistant questions to answer. At most one open refresh item per page.seo_list_roadmap({ status: "ignored" })— what the reconcile hid, withhiddenReason.seo_roadmap_create_brief({ itemId })— the Plan row's "Create brief" button. Idempotent (outcome"exists" if a brief is already there). Returns{ briefId, operationId, outcome }; pollget_job(operationId), then read it withseo_get_brief. Onseo_list_roadmap,brief.statusisnoneuntil a brief exists.- On a refresh item the brief is a refresh brief (
kind: "refresh",refreshPageUrl, and achanges[]edit plan of rewrite / add / keep / remove lines).seo_roadmap_edit({ itemId, action: "start" })on a refresh item does the same thing asseo_roadmap_create_brief. Nothing is rewritten yet. - Approving a refresh brief (
seo_approve_briefs) overwrites the existing article in place (blog articles only). Review thechanges[]with the user first; edit them withseo_edit_briefif needed. Useseo_approve_briefsfor refresh briefs:seo_write_articlereturns 400not_supported_for_refresh.seo_regenerate_briefrebuilds a refresh brief as a refresh brief (409brief_not_regenerablewhile writing or published).
Flow: Clusters -> Plan (ranked queue) -> Briefs -> Results. Every Plan row ends in a brief: seo_list_roadmap -> seo_roadmap_create_brief -> get_job -> seo_get_brief -> seo_approve_briefs -> get_job. seo_list_clusters (detail=medium) also returns demand per cluster: proven_demand (real impressions), covered (a page of ours already ranks top 3, skip it) or new (no data yet).
Overrides the reconcile always respects (use seo_roadmap_edit action):
restore— un-hide an item; it stays visible on later runs. Only applies to hidden/ignored items: on any other status it returns 409 (so completed work is never reopened).pin/unpin— freeze / unfreeze a suggested item's rank.remove— hide it and never re-create it (a keyword or topic the user does not want).
CLI: pking seo search-data (status), pking seo search-data reconcile, pking seo roadmap --type refresh, pking seo roadmap --status ignored, pking seo roadmap brief <id> [--no-wait] (create the brief), pking seo roadmap edit <id> --action <restore|pin|unpin|remove|start> [--no-wait]; pking seo clusters shows the demand tag.
Cluster / keyword / roadmap housekeeping (optional)
These are correction/cleanup tools, not steps in the main pipeline — reach for them when the auto-generated set needs fixing rather than running the numbered Procedure again.
Cluster housekeeping (approve/reject/bulk-approve/bulk-reject are already in the main flow at step 6):
seo_edit_cluster({ clusterId, name?, pillarKeyword?, description?, keywordsMeta?, contentMix?, briefAssignments? })— rename/redescribe, or change membership.keywordsMetais a full-REPLACEMENT array, not a delta — fetch the current array viaseo_list_clusters({ detail: "full" }), mutate it, and send the whole thing back (see Pitfalls).seo_delete_cluster({ clusterId, confirm: true })— hard delete; cascades to the cluster's roadmap items and their briefs. Not undoable — preferseo_reject_cluster(reversible viaseo_restore_cluster) unless the cascade is actually wanted.seo_merge_clusters({ clusterId, sourceClusterId })— reassigns the source cluster's keywords, roadmap items (and their briefs), and linked blog articles onto the targetclusterId, then deletes the source. Not undoable. The target's ownkeywordsMetaisn't updated by the merge — follow up withseo_edit_clusterif it needs to include the newly reassigned keywords.seo_restore_cluster({ clusterId })— undoseo_reject_cluster, back topending_review.seo_unapprove_cluster({ clusterId })— revertapproved→pending_review. Fails once briefs already exist for the cluster.
Keyword housekeeping:
seo_edit_keyword({ keywordId, intent?, userTags?, priority?, excludedFromClustering? })— override intent/tags/priority for one keyword, or hold it out of the next clustering pass. At least one field required.seo_bulk_edit_keywords({ keywordIds, intent?, userTags?, priority?, excludedFromClustering?, confirm: true })— same, across many IDs in one call.seo_delete_keyword({ keywordId, confirm: true })/seo_bulk_delete_keywords({ keywordIds, confirm: true })— soft-delete (setsdeletedAt; filtered out of every query, but recoverable).seo_restore_keyword({ keywordIds })— undo a soft-delete. IDs that aren't currently deleted are silent no-ops.seo_list_keyword_ids({ ...same filters as seo_list_keywords, e.g. relevanceMax, hasTag })— unpaginated ID export, no keyword data attached. Feed the result straight into the two bulk tools above.seo_tag_geography({ dryRun?, retag? })— unrelated to SEO market/language settings despite the similarly-general name: async, classifies each keyword's geographic search intent and writes ageo:*userTag (geo:us-mn,geo:us,geo:none,geo:ambiguous). RundryRun: truefirst to reviewtagCounts/ambiguousKeywordsbefore committing (idempotent by default — already-tagged keywords are skipped unlessretag: true). Then select withseo_list_keywords/seo_list_keyword_idsviahasTag.
Roadmap surface — seo_generate_roadmap (step 7 in the Procedure) is the only generation entry point; everything below is read/edit/delete on items it already created:
seo_list_roadmap({ status?, type?, priority?, clusterId?, detail? })— list items;status∈ suggested/in_progress/completed/ignored (ignored items are hidden unless you ask for them);type∈ blog_article/side_page/tool/refresh.seo_roadmap_get({ itemId, detail? })— single item, full detail by default.seo_roadmap_edit({ itemId, action?, status? })— change status, or override the search-data reconcile (see "Roadmap from real search data" below).action∈ ignore/restore/start/complete/pin/unpin/remove. There is no title or priority field.seo_roadmap_delete({ itemId, confirm: true })— permanent delete, irreversible.seo_roadmap_stats({ brandId? })— completed/in-progress/suggested counts, plus a dashboard link.get_seo_roadmap({ status?, detail? })— a legacy duplicate, registered inblog.tsrather thanseo/roadmap.ts: it combinesseo_list_roadmap+seo_roadmap_statsinto one call with a slightly different item shape (flattype/keywordfields instead ofprimaryKeywords). Prefer the two split tools above; only reach for this one if an existing integration already depends on its response shape.
CLI fast path
If a shell is available, the same flow via pking:
pking seo seeds "AI content generation" "social scheduling" "brand voice"
pking seo generate # expand seeds (long-running)
pking seo categorize
pking seo cluster
pking seo clusters list
pking seo clusters approve <clusterId>
pking seo clusters bulk-approve <clusterId1> <clusterId2>
pking seo briefs generate --cluster <clusterId> [--cluster <clusterId2> ...] # generates the roadmap + briefs (L3) for approved cluster(s)
pking seo roadmap [--cluster <clusterId>] # READ-ONLY: lists roadmap items, optionally filtered by cluster
pking seo briefs list [--status pending_review]
pking seo briefs view <briefId>
pking seo briefs regenerate <briefId>
pking seo briefs approve <briefId>
pking seo briefs bulk-approve <briefId1> <briefId2>
pking seo write <roadmapItemId> [--hero-image] [--voice <id>] [--no-wait] [--json]
pking seo gap
pking seo competitor --domain <competitorDomain>
pking seo publish --article-id <articleId> [--publication <id>] [--schedule <iso>]
pking seo stats
CTA auto-assignment is MCP/dashboard-only — there is no pking blogs auto-assign-cta command. Use the seo_auto_assign_cta({ blogIds }) MCP tool instead (blogIds is an array of up to 50 blog IDs, or the literal "all").
Brief edits (seo_edit_brief's structured briefData replacement) have no CLI command — do them via the MCP tool or the PostKing dashboard's brief editor. pking seo roadmap edit <id> supports only --action <ignore|restore|start|complete|pin|unpin|remove> and --status <suggested|ignored|in_progress|completed> (no --title/--priority). pking seo clusters approve <id> and pking seo briefs approve <briefId> each take a single id — use bulk-approve/bulk-reject (clusters) for multiple ids at once.
CLI-only granular operations (no direct MCP tool equivalent — useful when composing your own pipeline):
- List scored keywords:
pking seo keywords list --json— the full scored keyword list (keyword, volume, KD, intent). Pipe a filtered subset into CSV clustering below. - Cluster from a CSV:
pking seo clusters generate --csv keywords.csv --brand <brandId>(or--brand <brandId>alone to use the brand's existing keywords). Async — the CLI polls automatically and prints{clustersCreated, clustersMerged, keywordsAssigned}; pass--no-waitfor anoperationIdand poll manually withpking jobs list. CSV needs a header row with a requiredkeywordcolumn; optionalvolume,kd,intentcolumns improve cluster quality:
Other flags:keyword,volume,kd,intent ai content generation,8100,42,informational seo content tool,2400,38,commercial--target-size <n>,--no-cannibal-check(skip dedup check),--json(full cluster objects). - Generate a single brief from agent-supplied data:
pking seo briefs generate-one --pillar "ai content generation" --supporting "ai blog writer,seo content tool" --type blog --research advanced --brand <brandId> --json— produces one content brief without a roadmap, from data you supply directly. Async, polls and prints the fullbriefDataJSON. Required:--pillar,--supporting(comma-list, no spaces),--type blog|comparison|tool|landing,--brand. Optional:--intent,--research advanced|fast(defaultadvanced;fastskips external research),--tier hub|spoke|side_page,--json,--no-wait. The returnedbriefDatais the same structure accepted byseo_edit_brief'sbriefDataparam.
Expected Output
- A populated content roadmap tied to keyword clusters.
- Draft articles scored against primary keywords, each with FAQ/definition-first GEO structure.
- A gap-analysis report listing uncovered high-value topics (if run).
- At least one published or scheduled article (if credits + free tier allow).
Pitfalls
INSUFFICIENT_CREDITS— surface thecheckoutUrlfrom the error envelope and stop, or hand off to thepostking-getting-startedskill's billing section (billing_list_packs→billing_topup).FREE_CAP_REACHEDonseo_publish_article— same treatment asINSUFFICIENT_CREDITS.RATE_LIMITED— back off using theretryAftervalue from the error envelope before retrying.UNAUTHORIZED— re-run the login flow described in thepostking-getting-startedskill.clusterMetamissing/empty on a cluster — either no competitors are registered on the brand yet (see thepostking-competitor-intelskill), or the cluster simply doesn't overlap a known rival's keyword footprint (expected for purely core/adjacent clusters).seo_generate_keywords({ language })pointed at a language with no configured SEO market fails the async job (check viaget_job) with[keyword-pull] brand <id> has no configured SEO market for language "<lang>"— callseo_set_market_settingsfirst (seo_estimate_research_costwill have shownpairCount: 0for that language). Omittinglanguageentirely on a brand whose markets already span more than one language fails the same way with... has markets spanning N languages ... a keyword pull must specify opts.languageCode— passlanguageexplicitly in that case.seo_edit_cluster'skeywordsMetais a full REPLACEMENT array, not a delta — sending a partial list silently drops the missing members from the cluster. Fetch the current array viaseo_list_clusters({ detail: "full" }), mutate it, then send the whole thing back.- Calling
seo_edit_briefwith a free-textinstructionsparam — it's silently ignored; the inner route only acceptsbriefData(structured) and/orstatus. seo_generate_side_pagerejects aparentSlugparam name — the tool expectsslugfor the parent landing page.pking lp side generate <slug> --type <type>is currently broken — the inner route ignores the<slug>path param and requiresname/slug/landingPageIdin the request body, so it always 400s. Use the MCP toolgenerate_side_page({ slug, key, sidePageType })instead (sidePageType∈landing|text|comparison).
Verification
seo_list_clusters({ brandId, detail: "short" })should return without error and reflect any clusters just created.seo_list_briefs({ brandId })should list briefs whose count matches the roadmap items just generated.seo_list_results({ brandId })should show the article afterseo_write_articlecompletes.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| slack-gif-creatoranthropics/skillsKnowledge and utilities for creating animated GIFs optimized for Slack. Provides constraints, validation tools, and animation concepts. Use when users request animated GIFs for Slack like "make me a GIF of X doing Y for Slack." | Stars: 197k | Updated | Claude CodeCodex | Verified | Install |
| ponytailDietrichGebert/ponytailLazy senior dev mode for any coding task: the smallest change that fully solves it, and a reply a busy human understands in one read. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailHarvest every shortcut: comment into one debt ledger, so deferrals get tracked instead of forgotten. One-shot report. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |
| ponytail-debtDietrichGebert/ponytailList every `shortcut:` comment (and older `ponytail:` ones) as a debt ledger. One-shot report, changes nothing. Use for "ponytail debt", "what did ponytail defer", "list the shortcuts", /ponytail-debt. | Stars: 194k | Updated | Claude CodeCodex | Verified | Install |