[External Plugin]: pm-skills

[Avyayalaya]

Plugin name

pm-skills

Short description

MCP server and 12 skills for product analysis, specification writing, strategy, narrative, and stakeholder alignment. Run competitive analyses, write zero-question specs, design metrics, and produce executive-ready documents directly from Copilot.

GitHub repository

Avyayalaya/pm-skills-arsenal

Plugin path inside the repository

No response

Immutable ref to review

refs/tags/v2.1.0

Version

2.1.0

License identifier

MIT

Author name

Parth Sangani

Author URL

https://github.com/Avyayalaya

Homepage URL

No response

Keywords

product-management
competitive-analysis
specification-writing
product-strategy
go-to-market
metrics
discovery-research
narrative
stakeholder-alignment
mcp

Additional notes for reviewers

What's included

• MCP server (pm-skills): 5 tools — list_skills, get_skill, list_agents, get_benchmark, run_skill
• 12 skills (~1,000–1,300 lines each): competitive-market-analysis, discovery-research, problem-framing, specification-writing, metric-design-experimentation, product-strategy, go-to-market-strategy, pricing-packaging, executive-writing, narrative-building, multi-channel-publishing, stakeholder-alignment
• composes_with: frontmatter on the 5 highest-traffic skills declaring cross-package routing relationships to microsoft/azure-skills siblings. Companion guide at docs/composition.md.
• PostToolUse telemetry hooks (cross-platform bash + PowerShell; local-only JSONL append; no remote network calls)

Why these aren't "duplicate model strengths"

Frontier models can write a passable spec or summarize a market. These skills encode specific PM methodology the model doesn't have built-in: outcome-first spec structure with binary-testable acceptance criteria, 7 Powers framework for moat assessment, Goodhart's Law countermeasures in metric design, evidence tiering with H/M/L confidence, etc. Each skill is 1,000–1,300 lines of framework + decision rubric + failure modes. The uplift is methodological, not stylistic.

Benchmark — read with this lens

The repo cites 98/105 against a 7-dimension rubric vs Anthropic's PM Skill at 81/105 and a no-skill Claude baseline at 47/105. This is a self-administered benchmark — same author wrote the rubric AND scored every output across 5 prompts × 3 conditions. The rubric and all 15 raw outputs are published at benchmark/ for independent re-scoring.

Path note

The plugin.json sits at the repository root (not under .github/plugins/... like azure-skills). apm.yml declares source: ./. The Plugin path field is left blank per the form's "optional when at root" rule.

Submission compliance

Public GitHub repo (confirmed via gh repo view), MIT license, immutable ref (v2.1.0 tag pushed before this submission), no paid services, no RAI violations.

Submission checklist

• The plugin lives in a public GitHub repository.
• The ref I provided is an immutable release tag or full 40-character commit SHA, not a branch.
• This submission follows this repository's contribution, security, and responsible AI policies.
• This plugin is not already listed in the Awesome Copilot marketplace.

[github-actions]

🟡 Contributor Reputation Check: MEDIUM risk

Check	Risk
Profile	MEDIUM
Credential audit	NONE

Maintainers: please review this contributor before merging.
See the workflow run for full details.
Automated check powered by AGT.

[github-actions]

✅ External plugin intake passed

This submission passed automated intake validation and is ready for maintainer review.

• Plugin: pm-skills
• Repository: https://github.com/Avyayalaya/pm-skills-arsenal
• Ref: refs/tags/v2.1.0
• Keywords: product-management, competitive-analysis, specification-writing, product-strategy, go-to-market, metrics, discovery-research, narrative, stakeholder-alignment, mcp

Canonical external.json payload

{
  "name": "pm-skills",
  "description": "MCP server and 12 skills for product analysis, specification writing, strategy, narrative, and stakeholder alignment. Run competitive analyses, write zero-question specs, design metrics, and produce executive-ready documents directly from Copilot.",
  "version": "2.1.0",
  "author": {
    "name": "Parth Sangani",
    "url": "https://github.com/Avyayalaya"
  },
  "repository": "https://github.com/Avyayalaya/pm-skills-arsenal",
  "license": "MIT",
  "keywords": [
    "product-management",
    "competitive-analysis",
    "specification-writing",
    "product-strategy",
    "go-to-market",
    "metrics",
    "discovery-research",
    "narrative",
    "stakeholder-alignment",
    "mcp"
  ],
  "source": {
    "source": "github",
    "repo": "Avyayalaya/pm-skills-arsenal",
    "ref": "refs/tags/v2.1.0"
  }
}

Reviewer notes

What's included

• MCP server (pm-skills): 5 tools — list_skills, get_skill, list_agents, get_benchmark, run_skill
• 12 skills (~1,000–1,300 lines each): competitive-market-analysis, discovery-research, problem-framing, specification-writing, metric-design-experimentation, product-strategy, go-to-market-strategy, pricing-packaging, executive-writing, narrative-building, multi-channel-publishing, stakeholder-alignment
• composes_with: frontmatter on the 5 highest-traffic skills declaring cross-package routing relationships to microsoft/azure-skills siblings. Companion guide at docs/composition.md.
• PostToolUse telemetry hooks (cross-platform bash + PowerShell; local-only JSONL append; no remote network calls)

Why these aren't "duplicate model strengths"

Frontier models can write a passable spec or summarize a market. These skills encode specific PM methodology the model doesn't have built-in: outcome-first spec structure with binary-testable acceptance criteria, 7 Powers framework for moat assessment, Goodhart's Law countermeasures in metric design, evidence tiering with H/M/L confidence, etc. Each skill is 1,000–1,300 lines of framework + decision rubric + failure modes. The uplift is methodological, not stylistic.

Benchmark — read with this lens

The repo cites 98/105 against a 7-dimension rubric vs Anthropic's PM Skill at 81/105 and a no-skill Claude baseline at 47/105. This is a self-administered benchmark — same author wrote the rubric AND scored every output across 5 prompts × 3 conditions. The rubric and all 15 raw outputs are published at benchmark/ for independent re-scoring.

Path note

The plugin.json sits at the repository root (not under .github/plugins/... like azure-skills). apm.yml declares source: ./. The Plugin path field is left blank per the form's "optional when at root" rule.

Submission compliance

Public GitHub repo (confirmed via gh repo view), MIT license, immutable ref (v2.1.0 tag pushed before this submission), no paid services, no RAI violations.

[github-actions]

🟡 Contributor Reputation Check: MEDIUM risk

Check	Risk
Profile	MEDIUM
Credential audit	NONE

Maintainers: please review this contributor before merging.
See the workflow run for full details.
Automated check powered by AGT.

[aaronpowell]

/approve

[github-actions]

✅ External plugin approved

A maintainer approved pm-skills, and the submission issue has been closed.

• Catalog action: added
• Source repository: Avyayalaya/pm-skills-arsenal
• PR against staged: [external-plugin] Add pm-skills #1770