MCP HubMCP Hub
SKILL·C9DFF1

competitive-intel-watch

deanpeters
Mis à jour 6 days ago
5,952
730
5,952
Voir sur GitHub
Autreai

À propos

Cette compétence effectue une surveillance automatisée des écarts par rapport à un instantané concurrentiel antérieur, en signalant uniquement les changements significatifs avec des preuves citées et des indicateurs pour la mise à jour des battle-cards. Elle est conçue pour fonctionner de manière autonome selon un planning, en filtrant le bruit pour fournir des renseignements exploitables. Utilisez-la pour maintenir à jour les actifs concurrentiels sans régénération manuelle hebdomadaire des recherches.

Installation rapide

Claude Code

Recommandé
Principal
npx skills add deanpeters/Product-Manager-Skills -a claude-code
Commande PluginAlternatif
/plugin add https://github.com/deanpeters/Product-Manager-Skills
Git CloneAlternatif
git clone https://github.com/deanpeters/Product-Manager-Skills.git ~/.claude/skills/competitive-intel-watch

Copiez et collez cette commande dans Claude Code pour installer cette compétence

Documentation

Competitive Intel Watch

Purpose

Monitor a competitive landscape for material shifts since the last run. Diff the world against the previous snapshot; report only what changed, with evidence; flag which downstream artifacts need updating. This is the skill that turns competitive research from a document into a cadence — the weekly SIGINT sweep and monthly OSINT digest from the fusion cadence live here. A watch reports change, not state: regenerating the same report weekly is theater, and "no material shifts this cycle" is a valid, useful result.

Input

Works best with: the previous Competitive Research Snapshot (pasted or attached) — the baseline this run diffs against — and the competitor list (defaults to those in the snapshot). Also useful: anything specific you're watching for this cycle, and a materiality bar adjustment if the default needs tightening or loosening.

Input supplied inline with the invocation — text after the skill name, a pasted context dump, or an appended ARGUMENTS: line — counts as answers already given. Use it against the question budget; don't re-ask.

Arriving empty-handed? That works too. With no prior snapshot, the skill falls back to baseline mode: it produces a first snapshot using the competitive-research-snapshot structure and stops — the delta value starts on run two.

Example invocation: Competitive intel watch — prior snapshot pasted below; this cycle I'm specifically watching for pricing moves. [snapshot]

Key Concepts

  • Governing protocol: honors the autonomous-investigation contract — question budget of 2 (this skill's tightest), search-plan gate, Fact/Inference/Assumption labels, Just Enough Mode, stable schema, 4-option Final Step.
  • Discipline mix: SIGINT first (site diffs, pricing pages, job posts — the freshest layer), with OSINT and HUMINT signals monthly and FININT on the quarterly pass — the fusion cadence in intelligence-collection-disciplines is this skill's operating rhythm.
  • The materiality bar. Report a change only if a sales rep, pricing owner, or roadmap owner would plausibly act on it: pricing/packaging changes, launches and deprecations, positioning shifts, leadership moves, funding or M&A, major customer wins/losses, credible roadmap signals. Below the bar: cosmetic site changes, routine content marketing, minor releases. Why it matters: a watch that cries wolf gets ignored by cycle three — the bar is what keeps the audience.
  • Delta discipline. Read the prior snapshot fully before searching; diff against it, never regenerate it. The empty changelog is a first-class outcome.
  • Update flags close the loop. Research is only done when it names the artifact it changes — each material shift maps to the battle card, positioning, pricing, or roadmap sections now stale.
  • Do-not-invent list: competitors, features, pricing, market share, customer wins, roadmap items, product claims. Every claimed change carries a URL and a date.
  • When NOT to use: no baseline exists and you want the full treatment → run competitive-research-snapshot first; the scope itself changed (new segment, pivot) → re-snapshot from scratch rather than diffing a stale scope.

Application

  1. Determine mode. Prior snapshot provided → delta mode. None → baseline mode: produce a snapshot per the competitive-research-snapshot schema and stop.
  2. Credit inline context, then ask only the unanswered questions (max 2):
    1. Do you have the previous snapshot, or should I create a baseline?
    2. Anything specific you're watching for this cycle? If unanswered, proceed: baseline mode if no snapshot, default materiality bar otherwise.
  3. Read the prior snapshot fully before searching. The diff target is the document, not your memory of the market.
  4. Show the 3-bullet search plan — what you'll check per competitor, source types (company sites, pricing pages, release notes, press, investor materials, credible news, review sites, job postings), how facts will be separated from inference. Continue unless revised.
  5. Sweep and filter through the materiality bar. When nothing clears it, say so plainly.
  6. Emit the schema below exactly — runs must be diffable.

Output schema (do not reorder)

# Competitive Watch Report

## 1. Run Header
**Scope (from prior snapshot):** | **Prior snapshot date:** | **This run date:** | **Competitors checked:**

## 2. Changelog (Material Shifts Only)
For each material shift:
### [Competitor] — [4 to 8 word change summary]
- **What changed:** [1-2 bullets, labeled Fact/Inference]
- **Evidence:** [URL, date]
- **So what:** [why it clears the materiality bar]
- **Confidence:** [high / medium / low]

If nothing cleared the bar: "No material shifts this cycle." List
anything on the watchlist for next run.

## 3. Update Flags
| Downstream artifact | Sections needing update | Driven by |
|---|---|---|
| Battle card | | |
| Positioning statement | | |
| Pricing/packaging analysis | | |
| Roadmap assumptions | | |
Only rows with real updates; omit the rest.

## 4. Watchlist for Next Run
- [Signals below the bar but trending]
- [Open questions this run could not resolve]

### Assumptions to Validate
- [Assumption 1] / [Assumption 2] / [Assumption 3]

A copy/paste fill-in version of this schema, with quality checks, lives in template.md.

Final Step (offer exactly 4 options)

  1. Update the battle card sections flagged above (battle-card-builder)
  2. Deep-dive the most significant change
  3. Produce the refreshed full snapshot (new baseline)
  4. Adjust the materiality bar or competitor list for next run

Accept 1, 2, 3, 4, 1 and 2, Verbose Mode, or a custom path. On a scheduled, unattended run, file the report and stop — the options wait for a human.

Examples

A changelog entry that clears the bar (fictional):

Ledgerline — mid-tier plan removed from pricing page

  • What changed: the $49 "Team" tier no longer appears; feature list redistributed upward — Fact (pricing page vs. archived version, Jul 2 vs. Jun 1)
  • Evidence: URL + archive diff, dated
  • So what: entry price effectively doubled; our "cheaper to start" talking point is now stronger, and their SMB churn may spike — clears the bar for both sales and pricing owners
  • Confidence: high

The empty changelog done right:

No material shifts this cycle. Below-the-bar activity logged for trend: [Competitor B] published three thought-leadership posts on compliance automation (watchlist: possible positioning shift if their product pages follow), and two senior-engineer job posts mention a language we haven't seen in their stack before (watchlist: TECHINT corroboration needed before this means anything).

See examples/sample.md for a complete worked run (fictional FSM-software market) that diffs against the competitive-research-snapshot example's baseline — including an assumption from that baseline getting confirmed by the diff. examples/sample-industrial.md shows the quarterly-cadence industrial version, where a top risk gets demoted and that's reported as material.

Common Pitfalls

  • Regeneration theater. Producing a fresh full report each run and calling it a watch. The reader's question is "what changed?" — answer only that.
  • Materiality inflation. Reporting blog posts and minor releases to seem productive. Every below-bar item reported costs credibility the real alerts will need later.
  • Fear of the empty changelog. Padding a quiet cycle with noise. "No material change" backed by a real sweep is exactly what a healthy watch produces most cycles.
  • Undated evidence. A change claim without both URL and date can't be verified or diffed next run. The date is half the evidence.
  • Diffing a stale scope. The market pivoted, you entered a new segment — and the watch keeps diffing the old frame. Re-baseline when the scope changes; say so in the run header.
  • Orphaned intelligence. A changelog with no update flags. If no artifact needs updating, the shift probably didn't clear the bar — flags are how research becomes action.

References

Dépôt GitHub

deanpeters/Product-Manager-Skills
Chemin: skills/competitive-intel-watch
0
ai-agentsai-product-managementclaude-skillspm-frameworksproduct-management
FAQ

Questions fréquentes

Qu’est-ce que le Skill competitive-intel-watch ?

competitive-intel-watch est un Skill Claude créé par deanpeters. Un Skill regroupe des instructions et des ressources que Claude charge à la demande pour effectuer des tâches liées à competitive-intel-watch sans consigne supplémentaire.

Comment installer competitive-intel-watch ?

Utilisez les commandes d’installation de cette page : ajoutez competitive-intel-watch à Claude Code comme plugin ou clonez son dépôt dans votre dossier skills, puis redémarrez Claude pour charger le Skill.

À quelle catégorie appartient competitive-intel-watch ?

competitive-intel-watch appartient à la catégorie Autre.

competitive-intel-watch est-il gratuit ?

Oui. competitive-intel-watch est référencé sur AIMCP et son installation est gratuite.

Compétences associées

llamaguard
Autre

LlamaGuard est le modèle de Meta, doté de 7 à 8 milliards de paramètres, conçu pour modérer les entrées et sorties des LLM selon six catégories de sécurité comme la violence et les discours haineux. Il offre une précision de 94 à 95 % et peut être déployé avec vLLM, Hugging Face ou Amazon SageMaker. Utilisez cette compétence pour intégrer facilement le filtrage de contenu et des garde-fous de sécurité dans vos applications d'IA.

Voir la compétence
cost-optimization
Autre

Cette compétence de Claude aide les développeurs à optimiser les coûts du cloud grâce au redimensionnement des ressources, aux stratégies d'étiquetage et à l'analyse des dépenses. Elle fournit un cadre pour réduire les dépenses cloud et mettre en œuvre une gouvernance des coûts sur AWS, Azure et GCP. Utilisez-la lorsque vous devez analyser les coûts d'infrastructure, redimensionner les ressources ou respecter des contraintes budgétaires.

Voir la compétence
sports-betting-analyzer
Autre

Cette compétence Claude analyse les marchés des paris sportifs, incluant les spreads, les over/under et les paris spéciaux, en examinant les tendances historiques et les statistiques situationnelles pour identifier les paris à valeur ajoutée. Elle fournit une sortie en markdown structuré avec des recommandations actionnables à des fins éducatives. Les développeurs doivent l'utiliser pour des outils d'analyse de paris sportifs tout en notant qu'elle est conçue uniquement pour le divertissement et l'éducation.

Voir la compétence
quantizing-models-bitsandbytes
Autre

Cette compétence quantifie les LLMs en précision 8 bits ou 4 bits à l'aide de bitsandbytes, permettant une réduction de 50 à 75 % de la mémoire utilisée avec une perte de précision minime. Elle est idéale pour exécuter des modèles plus volumineux sur une mémoire GPU limitée ou pour accélérer l'inférence, prenant en charge des formats comme INT8, NF4 et FP4. La compétence s'intègre à HuggingFace Transformers et permet l'entraînement QLoRA ainsi que l'utilisation d'optimiseurs en 8 bits.

Voir la compétence