vendor-sklls-gap

name	vendor-sklls-gap
description	Fill vendor skill gaps when a built-in or custom skill reaches for the wrong tool or misses a vendor-specific workflow. Use proactively when a skill starts using the wrong integration, when the user asks to use a specific vendor workflow, or when invoked as /vendor-sklls-gap.

vendor-sklls-gap

Use this skill to patch gaps in vendor-oriented skills. Keep it organized by vendor/tool so more sections can be added over time.

Do not explain what you're about to do. Just run the commands.

Nia

When the user explicitly says ask nia, use /nia, or otherwise asks for Nia, do not use nia-docs as a substitute. Use the actual nia skill / workflow instead.

Only use npx nia-docs when the task is specifically about browsing a documentation site as a virtual filesystem and there is not a better vendor-specific path.

Nia docs browsing

Use npx nia-docs to browse a documentation site as a virtual filesystem. You can ls, grep, cat, and pipe — standard bash.

How to query docs

1. Index at the most specific URL you can

Top-level sites crawl shallowly. Subdirectories crawl deeply and give you the real content.

# Index the specific section you need
npx nia-docs https://docs.aws.amazon.com/AmazonECS/latest/developerguide/ -c "ls"
# ✔ Indexed docs.aws.amazon.com (414 pages)

If you don't know the subdirectory, start at the top level and navigate down:

npx nia-docs https://docs.aws.amazon.com/ -c "ls | grep -i ecs"
# → AmazonECS

npx nia-docs https://docs.aws.amazon.com/ -c "ls AmazonECS/latest"
# → APIReference / developerguide

Then re-index at the subdirectory for full depth.

2. Find files

# Find files by name
npx nia-docs <url> -c "ls | grep -i <keyword>"

# Grep inside files
npx nia-docs <url> -c "grep -rl '<pattern>'"

3. Read content

# Full file
npx nia-docs <url> -c "cat getting-started.html"

# Control output size
npx nia-docs <url> -c "cat getting-started.html" 2>&1 | head -150

4. Cache

After the first run the site is cached — subsequent calls are instant. No need to re-index.

Example: finding ECS Express Mode docs

# 1. Start at top level, find the ECS directory
npx nia-docs https://docs.aws.amazon.com/ -c "ls | grep -i ecs"
# → AmazonECS

# 2. Navigate down
npx nia-docs https://docs.aws.amazon.com/ -c "ls AmazonECS/latest"
# → APIReference / developerguide

# 3. Top-level crawl is shallow — only 2 files in the subdir
#    Re-index at the subdirectory for full depth (414 pages)
npx nia-docs https://docs.aws.amazon.com/AmazonECS/latest/developerguide/ -c "ls | grep -i express"
# → express-service-getting-started.html
# → express-service-create-full.html
# → express-service-work.html
# → ... (10 files)

# 4. Read the guide
npx nia-docs https://docs.aws.amazon.com/AmazonECS/latest/developerguide/ -c "cat express-service-getting-started.html"

Common documentation URLs

Service	URL
AWS ECS	https://docs.aws.amazon.com/AmazonECS/latest/developerguide/
AWS ECR	https://docs.aws.amazon.com/AmazonECR/latest/userguide/
AWS IAM	https://docs.aws.amazon.com/IAM/latest/UserGuide/
AWS App Runner	https://docs.aws.amazon.com/apprunner/latest/dg/
FastMCP	https://gofastmcp.com/
Vercel	https://vercel.com/docs/
Next.js	https://nextjs.org/docs/
Stripe	https://docs.stripe.com/
Claude Platform	https://platform.claude.com/docs/

Do not

• Do not run nia-docs interactively — always use -c "<command>".
• Do not skip the subdirectory step — top-level URLs give you a shallow index (e.g., 2 files instead of 414). Always navigate to the specific docs section and re-index there.
• Do not use search as a nia-docs command — it doesn't exist. Use ls, grep, cat.
• Do not forget 2>&1 — pipe stderr too, otherwise you miss the loading/indexing output.
• Do not forget timeout — set timeout: 120000 on Bash calls. First-time indexing can take 30-60s.