Tre defektklasser av samme slag: tall som ser sourcet ut, men ikke er det.
Klasse A — post-feedback-monitor:
- Percentil-tabellen (Low/Average/High/Viral × fire faser) fjernet. Ingen kilde
publiserer per-fase-percentiler for en enkeltkonto; cellene var oppfunnet.
Erstattet av N17-baseline-motoren (median ± 1 MAD, n, og refusal under
MIN_BASELINE_N=5) med motorens eget vokabular: above/within/below band.
- Velocity Score fjernet i sin helhet, inkl. fase-multiplikatorene (5,0x/3,0x/
1,5x/1,0x/0,5x). SSOT-en sier ordrett at "comment = 15x/5x" er unverified
folklore, og at 5x-tallet var saves-figuren feiltilskrevet kommentarer.
Erstattet av rå tellinger + engagement rate slik csv-parser.ts definerer den.
- Output-malen har nå en eksplisitt refusal-gren. Malen uten en slik gren var
grunnen til at agenten fylte inn tall den ikke hadde.
- To folklore-multiplikatorer i Principles ("5x the impact", "worth 15 likes").
Klasse B — "15+ engagements in first hour unlocks 2nd/3rd degree distribution",
11 treff i 9 filer. SSOT-en sier "Directional, not a fixed threshold". Påstanden
overlevde både hardening-gaten og kald-review (log.md:1099 sjekket ~70%-
misattribusjonen, ikke terskelen).
Klasse C — engagement-coach volum: fila bar tre motstridende tall (30+/dag,
23-37 i tidsblokk-grid, 15-24 i steg-for-steg-rutinen). Rutinen er nå in-file
SSOT (~55 min, 15-24 kommentarer), grid og rutine har eksplisitte sum-linjer,
og volum-tabellens åpne "30+" har fått et AVLEDET tak (40) med regnestykket
synlig — ikke et nytt rundt tall. Uverifiserbar superlativ ("110K followers,
#2 global creator") fjernet.
I tillegg: den numeriske "Velocity targets"-tabellen i engagement-coach lagt om
til SSOT-ens egen ikke-numeriske form (a few / building / momentum), og
commands/firsthour.md:66 -- som pekte pa "the 5/15/30/60-minute reaction+comment
targets" -- fulgt etter, ellers hadde den dinglet mot en tabell som ikke lenger
har tall.
docs/hardening/log.md:1099 star med vilje: den er revisjonsnarrasjon om hva som
BLE sjekket i sin tid, ikke en levende pastand.
Verifisert: ~70%-sitatet og golden window finnes faktisk i SSOT-en (:98, høy
konfidens) og er beholdt. Alle ti suiter grønne, floors uendret.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_017Pwb1oWLqKB2oBJHSoWNcy
Del 1 — scaffold-bånd (KTG-beslutning: behold målbåndet, utvid komponentene).
Komponentsummen var 960–1 640 mot målbåndet 1 200–1 800: en skjelett-konform
draft kunne lande under gulvet og nådde aldri taket. Nye bånd — Context
250–350, Insight 550–850, Implication 250–350 (Hook 110–140 og CTA 50–100
uendret) — summerer til 1 210–1 790, altså INNI 1 200–1 800. Skjelett-konform
er nå gate-konform per konstruksjon.
Målbåndet 1 200–1 800 er uendret overalt; hooks/prompts/content-quality-gate.md
(den kanoniske gaten) er ikke rørt. Alternativet — å heve taket per AuthoredUp
D-4 — ville krevd 25 filer / 40 linjer inkl. gate-prompten, begge skills,
quality-scorecard, config-malen og brain-fiksturen.
Skjelettet fantes i seks kopier, alle rettet på KTG-go: commands/post.md,
commands/batch.md, commands/pipeline.md, references/engagement-frameworks.md,
agents/content-optimizer.md, skills/linkedin-content-creation/SKILL.md. De tre
siste lå utenfor planens scope, men to av dem var internt selvmotsigende på én
linje (overskrift 1 200–1 800 over komponenter som summerte til 960–1 640), og
engagement-frameworks.md er nettopp fila post.md/pipeline.md instruerer
modellen om å LESE for strukturen.
Nabofunn tatt med på KTG-go: post.md:95 ga «Personal stories → 1 000–1 400»,
som motsa post.md:136s egen gate. Kald-review R2a MAJOR — nå 1 200–1 800.
D-4 inn i kanonfila: ny «Post length»-seksjon i
references/algorithm-signals-reference.md med AuthoredUp-optimumet 1 301–2 500
(372 126 poster, sep 2025–feb 2026), merket single-vendor/ett vindu, med
eksplisitt note om at datapunktet gjør det shippede taket konservativt — ikke
feil — og at de to AuthoredUp-N-ene i fila (621K vs 372K) er ulike studier.
Kilde verifisert mot primærkilden, ikke overført fra planen.
Del 2 — ferskhet-rester (begge påstander verifisert mot LinkedIn Help):
- D-6 newsletter-strategy-guide.md: «5 000+ følgere» framstilt som terskel er
feil — «All LinkedIn members have access to create a newsletter on LinkedIn»
(a517914). Omskrevet til redaksjonell modenhetsvurdering (også i
Mistakes-tabellen og Bottom Line). E-post er ikke garantert: LinkedIn
de-dupliserer på tvers av kanaler — «if you receive an in-app or push
notification, you should not expect to also receive an email for the same
notification» (a517914). Bringer referansefila i tråd med newsletter.md:2384.
Fjernet samtidig den ukildede «Algorithm favors newsletters from established
creators» i den omskrevne blokka.
- D-7 first-comment-strategy.md: «pinned by default» er uverifisert og feil —
pinning er en eksplisitt forfatterhandling (a524166), og standard
kommentarsortering er algoritmisk. Lagt til «What this file does not claim»-
avsnitt som speiler kanonfilas «contested»/low-confidence-epistemikk.
Scope 3 (KTG-go, amend): references/engagement-frameworks.md har FIRE
skjeletter, ikke ett. To til brot gulvet i malbandet — Data-Driven Post
(1 050-1 400) og Contrarian Post (1 060-1 410) — og er lagt om til samme
komponentprofil som Standard, sum 1 210-1 790. Narrative Arc (1 350-1 500) la
allerede inni og star urort; alle tre har na en eksplisitt sum-linje. Fila er
den post.md:104 sender modellen til for «story structures», sa a sertifisere
den som fikset med to odelagte skjeletter igjen ville vaert usant.
SUPERSEDED og ikke gjeninnført: gammel B §S6 Del 2 pkt 1 (first-comment
−5/−10 %-tall, pods-eskalering, 360Brew-fotnote).
Verifisering: bånd-summen ligger inni målbåndet i alle seks kopier av
standard-skjelettet og i alle fire skjelettene i engagement-frameworks.md
(grep-bevis, 0 gjenværende 200-300/400-800) · D-6/D-7 omformulert (0 treff på
«5,000+ followers» / «pinned by default» / «inbox + email») · alle ti suiter
grønne, alle floors uendret: test-runner 270/0 (269 assertions >= floor 251) ·
trends 300/0 · analytics 202/0 · hooks 191/0 · brain 134/0 · editions 72/0 ·
render 63/0 · specifics-bank 45/0 · tests 35/0 · contract-gate 33/0.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LhF1H7ctT5Fk8KkoCQpe5n
81 load-bearing stier i 16 kommandofiler pinnet til ${CLAUDE_PLUGIN_ROOT}/.
Rene prefikser: 81 inn / 81 ut, ingen linje lagt til eller fjernet.
FASIT-VERIFISERING (premiss-sjekk mot BASELINE, ikke bare HEAD): planens
68 linjer / 17 filer er RIKTIG mot sin egen baseline 882f6ee. HEAD ga 73/18 —
altså ekte kode-drift siden baseline (newsletter.md 11→15, calendar.md 0→1),
ikke en telle-feil som i N19.
SCOPE UTVIDET ETTER GO: planens grep (references|hooks/scripts|scripts) har en
blindsone — skills/ assets/ config/ render/ agents/ commands/ er samme defekt-
klasse og allerede anerkjent som pin-verdig av mønsteret (audit.md:24 og
competitive.md:22 pinner skills/, batch.md:214 og pipeline.md:211 pinner
assets/). Under planens scope kunne dens EGEN verifisering ikke passere:
quick.md:229-230 ville stått upinnet rett over en pinnet linje, og setup.md
(3 load-bearing Read, 0 pins) var ikke i settet i det hele tatt.
Full defektklasse: 117 linjer / 19 filer → 81 pinnet, 36 bevisst urørt.
KLASSIFISERINGSREGEL (utledet av mønsteret, ikke oppfunnet):
- PIN = stien skal åpnes/kjøres/skrives — Read/Reference/See-pekere,
ref-fil-lister, Bash/node/npm, edit-mål.
- PROSA = stien står som sitat bak en påstand, eller som beskrivelse av
oppførsel/plassering. Ingen inline «(see …)» er pinnet noe sted i repoet
(0 av 25 filer) — den grensen er arvet, ikke satt her.
36 BEVISST URØRTE, tre klasser:
1. Inline sitat/attribusjon (17): carousel.md:37 · pipeline.md:70,75 ·
newsletter.md:999,1082,1211,1433 · report.md:273 · video.md:69 ·
react.md:115 · calendar.md:87 · profile.md:28,34 · firsthour.md:112 ·
monetize.md:343,492 · outreach.md:922
2. Beskrivelse av oppførsel/plassering (10): newsletter.md:50,1097,1205,1694,
1744,1944 · report.md:45 · linkedin.md:218 · import.md:121,200
3. newsletter.md fase-tabellens Tooling-kolonne (9): :100,102,105,106,111,112,
113,115,116 — tabellen er et register, og hvert steg pinner sitt eget
faktiske kall i brødteksten.
S13-LINTEN (M0 data-dir): alle assets/-stiene som ble pinnet står eksplisitt i
lintens NEGATIVE13-unntaksliste, og lintens egen kommentar (test-runner.sh:552)
sier at shipped read-only assets SKAL bære ${CLAUDE_PLUGIN_ROOT}.
VERIFISERING: 76 unike pinnede stier hentet ut og resolvet med cwd=/tmp —
75 finnes. Den ene som ikke gjør det, config/personas.local.md, er en
gitignored valgfri fallback («else personas.template.md») som var pinnet på
newsletter.md:205 før denne endringen. Re-grep: gjenstående 36 er utelukkende
prosa, listet over.
Alle ti suiter grønne, alle floors uendret: test-runner 270/0 (269 assertions
>= floor 251) · trends 300/0 · analytics 202/0 · hooks 191/0 · brain 134/0 ·
editions 72/0 · render 63/0 · specifics-bank 45/0 · tests 35/0 ·
contract-gate 33/0.
Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01M31cw43gqiSSDrKzcwUi7L
The last de-niche slice: recast the 10 sites where the vendor/sector beat
(Microsoft|Azure|Copilot|public sector) sat as the PRIVILEGED/default example,
varying each to a concrete cross-domain example instead of sterilizing
(plugin-is-domain-general — domain comes from user config, never hardcoded).
Recast (10): url-processing-templates (news worked-example Copilot->Figma),
opportunity-generation (3 headline examples + About block -> varied/ops persona),
profile (3 "good example" headlines/impact -> healthcare/e-commerce/support),
first-comment-strategy (drop "Microsoft" from research-paper example),
poll-strategy-guide (Copilot option -> generic AI assistants),
engagement-frameworks (1 of 3 direct-address audiences -> RevOps/SaaS),
setup (audience e.g. -> two varied examples), post (invocation e.g. -> SaaS pricing),
network-builder (tagline example -> ops/manufacturing),
video-scripter (2 filename slugs -> neutral topics).
Kept as false positives (would sterilize): content-angles.md (Public Sector is
1 of 6 balanced industry tables + Industry-Agnostic section), outreach.md
(Microsoft Build/Ignite/Azure UG = 3 of ~20 varied real conferences),
linkedin-growth-playbook (biographical fact in a real case study), the
Gemini/Tavily/Perplexity MCP tool-name examples, and the algorithm-signals
"Gemini provenance" SSOT citation. AI-as-topic kept (not a niche token; the
de-AI/AI-slop mechanic is the plugin's legit subject).
Gate scripts/test-runner.sh 87/0/0 (no lint touches these files yet; §17-guard
extension to content-planner is the deferred next step). 10 files, 26/26.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01RBMKqPSVbvSZHtQ4heM1UY
Final sub-pass of the S29 plugin-wide terminology scrub. The canonical reference file is
renamed and every functional pointer updated atomically in one commit. The file's in-file
title/headers were already FORM A-scrubbed in S29c (H1 reads "Content Angles"), so S29e is a
pure rename + pointer update — no FORM A remained in the file.
Rename: references/thought-leadership-angles.md -> references/content-angles.md (git mv).
Pointers updated (17 files, 29 occurrences) — token "thought-leadership-angles" -> "content-angles":
- references/ (2): ai-content-framework, glossary
- agents/ (7): content-repurposer, strategy-advisor, network-builder, content-planner,
trend-spotter, video-scripter, differentiation-checker
- commands/ (6): pipeline, video, post, competitive, react, batch
- skills/ (1): linkedin-content-creation/SKILL
- docs/ (1, forward-looking): integration-test-guide
Left URØRT per the standing S29 decision (history = honest record of a past state, not a
runtime load): CHANGELOG.md, docs/hardening/log.md, docs/hardening/plan.md. STATE.md untouched
here (rewritten at session end).
Verify: no thought-leadership-angles* file remains; references/content-angles.md present; zero
residual "thought-leadership-angles" in commands/agents/references/skills/integration-test-guide;
structure gate scripts/test-runner.sh 81/0/0 exit 0.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016qgzo6rxthw7KuxHjn5vyE
First sub-pass of the S29 plugin-wide terminology scrub: the banned brand phrase
"thought leadership" (FORM A) removed from the user-facing command surface. 15 in-file
phrase occurrences across 9 commands re-sourced to neutral wording (post, video, strategy,
batch, setup, react, outreach, onboarding, monetize) — e.g. "thought leadership content
creator" → "content creator", "Select Thought Leadership Angle" → "Select Content Angle",
"Build thought leadership & authority" → "Build authority & influence", onboarding guide
→ "LinkedIn Studio onboarding guide".
Scope (operator-set): FORM A only. FORM B ("thought leader" as role, incl. the buzzword
denylist at post.md:134) left untouched. The 11 filename pointers to
references/thought-leadership-angles.md deferred to S29e (atomic rename + pointer update);
competitive.md/pipeline.md carry only pointers and are untouched here. Pure history
(CHANGELOG, docs/**) left intact per "history → git".
Verify: assertive non-filename FORM-A in commands/ = NONE; FORM B count unchanged (15);
diff +15/-15 across 9 files (1:1); gate 81/0/0; counts 29/19/26/6 + v0.5.0 unchanged.
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_016qgzo6rxthw7KuxHjn5vyE
Wire the orphan differentiation-checker (#10) into the five short-form
creation commands (post/quick/react/carousel/video) as a De-AI /
Differentiation Gate at each command's quality-check step: confirm the
LinkedIn-named substance signals (personal substance, original thinking,
concrete specifics, genuine voice) + a soft engagement-bait check, and
delegate an originality pass to linkedin-studio:differentiation-checker
when the angle risks commodity content. Add Task to allowed-tools in
quick/react/carousel (post/video already had it from Step 13).
Extend (not duplicate) hooks/prompts/voice-guardian.md's AI-pattern
section with the same named signals from research/01 D8 + research/03 D4.
Runtime-loaded prompt — no compile-hooks.py, no hooks.json change
(verified: compile-hooks --check reports no drift).
Test: new hooks/scripts/__tests__/linkedin-content-filter.test.mjs pins
the content/non-content boundary the gate is scoped by (14 tests).
Full hook suite 76/76, structure lint 61/61.
Plan Step 14 (Wave 4 S2). Counts unchanged (26 commands / 19 agents).
[skip-docs]: tre-doc + version bump deferred to Step 21 per remediation plan.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Resolves the orphan-agent audit finding by the locked default: wire all, no deletions, so the agent count stays 19. Per agent, added Task to the target command's allowed-tools and a coherent 'subagent_type: linkedin-studio:<name>' delegation at a real point in the command's flow (not a token grep-match).
Wired (agents 1-9 of 11): video-scripter -> video.md (Step 4); content-optimizer -> post.md (Step 7 refinement) + ab-test.md (2a.4 optimized challenger); analytics-interpreter -> report.md (Step 7, report mode) + analyze.md (Step 2, interpret mode); content-planner -> batch.md (Step 2) + pipeline.md (Step 1); trend-spotter -> batch.md (Step 1) + pipeline.md (Step 1); network-builder -> outreach.md (Step 3a); strategy-advisor -> strategy.md (Step 3); voice-trainer -> setup.md (Step 3a); post-feedback-monitor -> calendar.md (publish action, 48h monitor).
Deferred to their dedicated steps: #10 differentiation-checker -> Step 14 (short-form de-AI gate), #11 engagement-coach -> Step 16 (first-hour command). Namespaced subagent_type form requires a session reload before the wired agents are invokable.
Verify: each of the 9 has >=1 invocation in commands/; structural lint 61/61 (counts 19/26/25/6 intact); agent-fixtures 35/35; hook tests 62/62. Three-doc + version reconciliation deferred to Step 21 per the locked plan [skip-docs].
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
BREAKING CHANGE: the marketplace slug, the agent namespace
(linkedin-studio:<agent>), and the runtime state-file path
(~/.claude/linkedin-studio.local.md) all change. Reinstall required;
existing state migrated in place (post metrics, streak, history preserved).
The /linkedin:* commands are unchanged — the command namespace is set
per-command in frontmatter and was always independent of the plugin slug.
Functionality is byte-identical to v2.4.0; this release is pure identity.
- dir + manifests: plugins/linkedin-studio + plugin.json + root marketplace.json
- agent namespace updated in commands/newsletter.md (only functional invoker)
- state path updated in 4 hook scripts + topic-rotation prompt + state template
- catch-all skill dir renamed skills/linkedin-studio (5 functional skills unchanged)
- docs + version bump to 3.0.0 across README badge, CHANGELOG, root README/CLAUDE.md
- historical records (CHANGELOG past entries, docs/ build artifacts,
config-audit v5.0.0 snapshots) intentionally retain the old slug
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
- Skip topic ask if user provided one with command invocation
- Auto-select strongest angle instead of 3-option AskUserQuestion menu
- Infer format from content type instead of interactive selection
- Present ONE draft without proactive alternative versions
- Replace refinement AskUserQuestion with plain text options
Net result: 0 AskUserQuestion calls in main flow (was 2).
User provides topic → gets ONE draft → clipboard-ready.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Adds Bash to allowed-tools and clipboard-helper.mjs auto-copy to:
post.md, quick.md, react.md, pipeline.md, first-post.md,
video.md, multiplatform.md, carousel.md.
Each command auto-copies the final post/caption text to clipboard
after presenting the draft. Replaces manual copy-paste instructions
in first-post.md (=== COPY FROM HERE ===) with auto-copy.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add /ultraresearch-local for structured research combining local codebase
analysis with external knowledge via parallel agent swarms. Produces research
briefs with triangulation, confidence ratings, and source quality assessment.
New command: /ultraresearch-local with modes --quick, --local, --external, --fg.
New agents: research-orchestrator (opus), docs-researcher, community-researcher,
security-researcher, contrarian-researcher, gemini-bridge (all sonnet).
New template: research-brief-template.md.
Integration: --research flag in /ultraplan-local accepts pre-built research
briefs (up to 3), enriches the interview and exploration phases. Planning
orchestrator cross-references brief findings during synthesis.
Design principle: Context Engineering — right information to right agent at
right time. Research briefs are structured artifacts in the pipeline:
ultraresearch → brief → ultraplan --research → plan → ultraexecute.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
