competitive-teardown
Skill0 starsChanged 53 days agoMITView on GitHub
Produces a structured teardown of a competitor product covering positioning, pricing, onboarding, core loop, gaps, what to steal, and what to avoid, based only on material the user provides or public pages they point to. Use when the user asks to analyze a competitor, do a competitive teardown, or compare their product against another company's.
11 other projects use a version of this.
Adopted by
Most recently taken up by
- Bellistri-Nick/pm-skillstheir own version, first seen Nov 15, 2026
- darkweb19/skillstheir own version, first seen Nov 8, 2026
- brittanyslay/marketing-skillstheir own version, first seen Nov 5, 2026
- tderoussel/agent-skillstheir own version, first seen Oct 20, 2026
- Bellistri-Nick/ProductManagerCompetitiveAnalysistheir own version, first seen Oct 13, 2026
- scayver/marketing-skillstheir own version, first seen Oct 13, 2026
Other versions11 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 11
- Hassaan146/claude-skills★ 18same name
- scayver/marketing-skills★ 8same name
- Bellistri-Nick/ProductManagerCompetitiveAnalysis★ 2same name
- Hussain-Elfalah/Agentic-Skills★ 2same name
- mahiunfiltered/claude-code-skills★ 1same name
- tderoussel/agent-skills★ 1same name
- andfrancot-alt/claude-skills★ 0same name
- Bellistri-Nick/pm-skills★ 0same name
- brittanyslay/marketing-skills★ 0same name
- darkweb19/skills★ 0same name
1 more not listed
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository4 other files it gives its agents.
- eval-rubric-writerSkill
- launch-checklistSkill
- prd-drafterSkill
- user-interview-synthesizerSkill
The file
We split this skill into the rules it gives an agent
10 rules. 0 are common, 10 are distinctive, and 0 are very specific to this project.
0
Common
Most of the 11 edited versions we read keep these.
10
Distinctive
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 16
Before drafting, check what material is available:
- Distinctive: line 21
If you have a browsing tool available and the user asked you to look something up, only visit pages the user named or that are one click …
- Distinctive: line 28
If the user names a competitor with no material and no URL, ask for at least a homepage and pricing page link, or ask them to paste what …
- Distinctive: line 37
Start with the sources list at the top and keep it updated as you use each source.
- Distinctive: line 47
Onboarding: only what was observed or shown in the material.
- Distinctive: line 52
Gaps: only from evidence (a missing feature visible in the material, a stated complaint if review sources were given).
- Distinctive: line 58
Every entry in these two sections must point back to a specific claim made earlier in the teardown.
- Distinctive: line 73
If the source material conflicts with something you recall generally about the competitor (a price that seems out of date, a feature you …
- Distinctive: line 79
If the user's own product is being compared, do not editorialize about which is better.
- Distinctive: line 83
If a section has no supporting material at all, write that section as "not covered" rather than deleting it.
“Common” is measured: kept by at least 6 in 10 of the 11 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Competitive teardown
The single rule that matters more than the template: every claim in the output must trace to something the user gave you or a page you were explicitly pointed at and actually read. Anything else is a guess, and a guess presented as a fact is worse than no teardown at all, because it will get repeated as if it were verified.
Step 1: Confirm what you have to work with
Before drafting, check what material is available:
- Screenshots, PDFs, or text the user pasted directly.
- Specific URLs the user gave you (their pricing page, their onboarding flow, a review site).
- If you have a browsing tool available and the user asked you to look something up, only visit pages the user named or that are one click from a page they named (for example, the pricing link from their homepage). Do not go searching the open web for claims to fill in gaps; that turns "based on what you gave me" into "based on what I found," which the user did not ask for and cannot verify.
If the user names a competitor with no material and no URL, ask for at least a homepage and pricing page link, or ask them to paste what they want covered. Do not proceed from general knowledge of the competitor alone. General knowledge of a product is frequently stale, and a teardown built on stale knowledge reads as current when it is not.
Step 2: Fill the template section by section
Use templates/teardown-template.md. Start with the sources list at the top and keep it updated as you use each source; the sources list is not decoration, it is what makes every later claim checkable.
For each section, write only what the source material supports:
- Positioning: their stated claim, not your read of their strategy.
- Pricing: what is written, tier by tier. Anything inferred (for
example, guessing what an "Enterprise, contact us" tier costs) gets
marked
[UNVERIFIED]inline, not stated as fact. - Onboarding: only what was observed or shown in the material. If no onboarding material was provided, write "not covered, no source material provided" rather than describing a generic onboarding flow.
- Core loop: describe it as steps a user repeats, inferred from what the product does per the source material, not a feature list.
- Gaps: only from evidence (a missing feature visible in the material, a stated complaint if review sources were given). If no gap evidence exists, say so rather than inventing a plausible-sounding weakness.
Step 3: Write "what to steal" and "what to avoid"
Every entry in these two sections must point back to a specific claim made earlier in the teardown. If you cannot point to the section and evidence that supports a "steal" or "avoid" entry, it does not belong in the output. These sections are the most likely place for unsupported opinion to creep in; hold them to the same sourcing bar as the rest.
Step 4: Close with what's unverified
The "Unverified or out of scope" section is not optional. List anything relevant that could not be confirmed: pricing not publicly listed, architecture, team size, funding, user counts. This section exists so the reader knows the boundary of what this teardown actually established.
Decision rules
- If the source material conflicts with something you recall generally about the competitor (a price that seems out of date, a feature you believed they didn't have), trust the source material and note the discrepancy rather than silently going with prior knowledge or silently going with the source. Flag it: "source states X; this may have changed since general knowledge of this product was last current."
- If the user's own product is being compared, do not editorialize about which is better. State what each does per the material and let the "what to steal" and "what to avoid" sections carry the judgment, explicitly and with evidence.
- If a section has no supporting material at all, write that section as "not covered" rather than deleting it. A missing section that is clearly marked is more useful than one that quietly disappeared.
Anti-patterns
- Presenting a guess about pricing, user counts, or architecture as a
stated fact instead of marking it
[UNVERIFIED]. - Writing "their onboarding is bad" or "their UX is confusing" without citing the specific screen, step, or quote that supports it.
- Filling gaps in the material from general or training knowledge of the competitor without flagging that the information may be outdated.
- Treating a single negative review as a confirmed product-wide gap without noting it is one data point.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

