devrel-competitor-analysis skill (samber): benchmark a competitor's developer relations

• Asks up to ten questions one at a time, skipping what you already stated.

  Questions 8 to 10 (date, one-off or standing asset, effort ceiling) cap the pass; a half-day ceiling deletes community, events and hiring signals.

  Interview

• Caps the set at 3 to 5 competitors with a role each and writes down who was excluded.

  The cap is the analysis's own convention, not an industry standard.

  Frame the peer set

• Freezes identical observations for every company, including yours.

  Symmetry is what makes the table possible.

  Fix the scorecard before collecting

• Collects eight surfaces in order of finding per analyst hour, dating every observation.

  Timed quickstart, docs mode coverage, forge release and response metrics, Q&A tags, content cadence, community, hiring signals, events.

  Network · Collect the signals

• Times each competitor's getting-started path and stops at the first paid step.

  SKILL.md calls this the comparison no scraped metric gives you.

  Runs code · Run the quickstart yourself

• Normalizes numbers, then compares artefact types and cadence rather than volume.

  Where no divisor is defensible it reports the structural fact instead.

  Normalize, then read structure and cadence

• Sorts gaps by impact per cost and gives each a verdict: close, ignore or counter.

  Close needs an owner and date; ignore needs a written reason; counter names the strength you lean on.

  Rank the gaps and assign verdicts

• Writes a six-section report and saves dated observations so the next pass is a diff.

  What changes because of this, Scorecard, Per-surface read, Time-to-first-success, Gap plan, Method and limits.

  Try: “benchmark our developer experience against Northwind and Kestrel”

  Benchmark report

Watch out for

• The quickstart timing can run third-party commands

  Step 4 has the agent follow each competitor's getting-started path with a timer, stopping at a paid step or a sales contact. Our advice: use a throwaway directory or container and read each command before it runs. github.com

• Fetched competitor pages are untrusted input

  SKILL.md says marketing pages, docs and READMEs may carry text aimed at agents, and that instructions inside a fetched page are data to report, never commands to follow. github.com

• It reads bulk data from public APIs

  SKILL.md prefers a documented API over page extraction and says to check a venue's terms. Our advice: confirm the terms of each site and API before you let the agent pull volume data. github.com

• It may keep notes in devrel-context.md

  SKILL.md tells the agent to store a few details about the current task or project, and to use a file called devrel-context.md when its memory lives in a file. Our advice: open that file before you commit or share the folder. github.com

• The plugin route installs the whole collection

  The README's plugin and extension routes add every skill in the repo; its git tree holds 93. The repo has no hooks, agents, commands or MCP server files, and 14 of its skills bundle a script. Our advice: use the skills CLI command for this skill if you want only it. github.com

Security advisories