Guide · Agent skill
How to use the Benchmark Methodology agent skill
Updated · written with AI from the project's own documentation and checked against our catalogue data
The benchmark-methodology skill gives an AI agent a consistent way to compare competing businesses, such as design or branding studios. It scores each competitor on nine weighted dimensions using a shared 1-5 scale tied to observable evidence. This guide explains what the skill does, how to install it, how to trigger it with plain requests, and where it fits and where it doesn't.
Benchmark Methodologyaffaan-m/ECC · MIT Download the skill Quick start
- 1Download the ZIP. Get the benchmark-methodology ZIP from this page. It contains agents/openai.yaml and SKILL.md.
- 2Install it. Unzip into ~/.claude/skills/ for personal use or into the project's .claude/skills/ folder, or upload the ZIP under Settings → Capabilities → Skills in the Claude apps.
- 3Prepare a tiered competitor set. Use the competitive-platform-analysis step to produce a scoped, tiered list of competitors.
- 4Write the client positioning brief. Name the strategic tension, its two axes, the differentiator and the brand balance so the tension dimension can be scored.
- 5Ask for a scored benchmark. Make a plain request that names the competitors and the brief, and asks for evidence-backed scores.
- 6Check the evidence. Confirm each score has a justification and source, and downgrade uncorroborated claims.
- 7Review the tension plot. Look at where competitors sit on the 2x2 and who occupies the client's target quadrant.
What is an agent skill and what does this one make the agent do?
An agent skill is a folder of instructions that an AI agent loads when a task matches it. This skill is named benchmark-methodology and is defined in a file called SKILL.md. Its purpose is to make comparisons repeatable: every competitor is judged on the same dimensions, with the same rubric, so scores can be compared fairly.
The skill scores nine dimensions: positioning clarity and distinctiveness (18%), brand voice (15%), visual identity and site craft (15%), service offer and packaging (12%), evidence and credibility (12%), enterprise-readiness (10%), thought leadership (8%), pricing transparency (5%), and the client's strategic tension (5%). The weights guide emphasis and sum to 100%. The skill deliberately does not produce a single blended score, because an average hides the differences that matter.
- Scores each competitor 1-5 on the first eight dimensions, with level descriptions from 1 (absent or generic) to 5 (category-defining).
- Scores both poles of the client's strategic tension separately and never averages them.
- Asks the agent to downgrade claims that have no corroboration, since site copy is marketing rather than proof.
- Produces a profile card per competitor with a score, one-line justification and source for each dimension.
Install the benchmark-methodology skill
You can install the skill in one of two ways. The first is to download the ZIP from this page and unzip it into your personal skills folder, ~/.claude/skills/. To make it available only inside one project, unzip it into that project's .claude/skills/ folder instead.
The second way is to upload the ZIP in the Claude apps. Open Settings, go to Capabilities, then choose Skills and upload the file there.
The package contains two files: agents/openai.yaml and SKILL.md. SKILL.md holds the methodology the agent follows.
- Personal install: unzip into ~/.claude/skills/
- Project install: unzip into the project's .claude/skills/
- Claude apps: Settings → Capabilities → Skills, then upload the ZIP
What you need before you start
The skill's listed requirements are nothing in particular. It does not run scripts and does not use the network. Its work happens inside the agent's conversation, using the material you provide and whatever research you or the agent gather.
The skill expects two inputs to be ready. The first is a scoped, tiered competitor set, which the description says comes from the competitive-platform-analysis skill. The second is a client positioning brief that names the strategic tension and its two axes. Without the brief, the tension dimension cannot be scored.
- A scoped, tiered list of competitors (the skill's description says to use it after competitive-platform-analysis)
- A client positioning brief with the strategic tension, differentiator and brand balance
- Access to each competitor's site, case studies, reviews and social channels, so that scores can be backed by sources
How to trigger the skill with natural requests
You do not need a special command. The skill's description tells the agent when it applies: after a tiered competitor set is ready and comparable, evidence-based scores are needed. Describe that situation in plain words and name the competitors and the client's tension.
A good request states the goal, the competitors and the brief. For example: score these three studios against our memorability versus credibility brief, using evidence from their sites and review directories. Asking for a single ranking with no evidence is the kind of request the skill is designed to avoid.
- Score these five competitors on the nine dimensions, using our positioning brief.
- Plot the competitors on our tension 2x2 and tell me who occupies the target quadrant.
- Write a profile card for each studio with a score, justification and source per dimension.
- Check whether our competitors' evidence scores hold up once self-reported claims are discounted.
Walk through a scoring session
The skill asks the agent to gather evidence in a fixed order, cheapest signal first. It starts with the competitor's own site, then case studies, review directories, LinkedIn, portfolio or craft platforms, and content channels. For credibility, it asks the agent to separate asserted claims (we delivered X) from proven ones (named, measurable, verifiable).
Each score must have a justification and a source. The skill also asks the agent to re-read all scores side by side before finishing, so that a 4 means the same thing for every competitor.
- Record the score, a one-line justification and the source link or screenshot for each dimension.
- Downgrade evidence and credibility scores for claims that nothing corroborates.
- Keep craft and credibility scored separately, so a plain site is not penalised for being plain.
- Report the tension plot separately from the dimension scores.
Safety notes: scripts and network access
The skill runs no scripts and uses no network connection, according to its listing. It is a set of written instructions. The agent may still browse or look up sources during the work, depending on the tools available in your environment, so review what the agent accesses for any client research.
The skill does not verify facts. It guides how to judge sources, but it does not check financial or market data. Treat any figures from competitors as claims to be confirmed.
- No scripts are run by this skill.
- No network use is listed for the skill itself.
- The skill does not check facts; verify financial or market figures yourself.
What people use it for
- A studio owner preparing a competitive review before pitching for SaaS or fintech work.
- A marketing strategist repositioning a brand who needs evidence-backed competitor scores instead of gut-feel rankings.
- A student learning structured competitive analysis who wants a worked example of consistent scoring.
- A team building profile cards that a competitive report can later assemble.
Frequently asked questions
How do I install the benchmark-methodology skill?
Download the ZIP from this page and unzip it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder for one project. In the Claude apps you can instead upload the ZIP under Settings → Capabilities → Skills.
Is the benchmark-methodology skill free to use?
The skill is published under the MIT licence, as listed on this page. This page does not state any price, so check the licence text in the source repository for the terms that apply.
What does the skill need to run?
Its listed requirements are nothing specific. It runs no scripts and uses no network. For good results you need a scoped, tiered competitor set and a client positioning brief that names the strategic tension.
Can I use it for a quick opinion on one competitor?
It is not designed for that. The method expects a scoped competitor set and multi-source evidence gathering, so a quick informal opinion on a single competitor is better done without it.
Does the skill give one overall score for each competitor?
No. It reports the nine dimension scores and the tension plot separately, because a weighted average hides the asymmetry that matters. The weights only guide emphasis.
How does this compare with other competitive analysis methods?
The listing does not compare it with other methods. Its distinctive features are the fixed 1-5 rubric, the separate scoring of both tension poles, and the downgrade of uncorroborated claims.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you run a studio or marketing function that needs a consistent, evidence-backed competitor review, or if you are learning structured competitive analysis. Pick something else if you need a fast opinion on one rival, if you have no positioning brief, or if you need verified financial or market data, since the skill does not check facts.
Download the skill