daymade__claude-code-skills/competitors-analysis/references/analysis_checklist.md

Competitor Analysis Checklist

Use this checklist before, during, and after any competitor analysis. The purpose is to keep repository evidence, market evidence, and judgment separate.

1. Scope And Storage

• Product or market scope is explicit.
• Competitor base directory is explicit: COMPETITORS_BASE="${COMPETITORS_BASE:-$HOME/workspace/competitors}".
• Product directory exists under $COMPETITORS_BASE/{product-slug}/.
• Repository directory uses the owner-repo convention.
• Any existing local clone is reused instead of cloning into a second path.

2. Discovery Checks

For broad searches, run more than one query:

gh search repos "primary keywords" --limit 30 --archived=false \
  --json fullName,url,description,stargazersCount,forksCount,openIssuesCount,language,pushedAt,updatedAt,defaultBranch

• Search queries are recorded.
• Candidate relevance is tied to the user's product scope.
• Broad search results are shortlisted before deep analysis.
• Forks, archived repos, and stale repos are identified rather than silently treated as first-class competitors.

3. Repository Preparation

repo="$COMPETITORS_BASE/{product-slug}/{owner-repo}"
test -d "$repo/.git"
git -C "$repo" remote -v
git -C "$repo" fetch --all --prune
git -C "$repo" log -1 --format='%H%x09%cI%x09%s'

• Remote URL is recorded.
• Latest local commit hash is recorded.
• Commit date is recorded.
• Default branch or current branch is recorded.
• Local changes, if any, are noted before pulling.

4. Source Reading

• README or docs are read for positioning.
• Config files are read for language, framework, scripts, and dependencies.
• Entry points are identified from config or file layout.
• Core implementation files are read directly.
• Tests or fixtures are checked when the competitor handles structured data.
• Changelog/releases are checked when the user asks for "latest".

5. Citation Checks

Use line-numbered reads before writing technical claims:

nl -ba package.json | sed -n '1,140p'
nl -ba src/main.ts | sed -n '1,220p'

• Versions cite config file lines.
• Feature claims cite README/docs and implementation lines when available.
• Parser/export/storage claims cite code lines.
• Market data cites GitHub/API/web source plus retrieval date.
• Each comparison-table value has a source cell.

6. Language Checks

Search the final report for unsupported language:

rg -n "(推测|可能|应该|大概|似乎|或许|未知|未披露|未公开|assume|probably|maybe)" profile.md

• No unsupported inference is presented as fact.
• Unknowns are written as 待验证 with a specific next check.
• Judgment is separated from repository facts.

7. Landscape Checks

For multi-competitor reports:

• Source register lists local path, remote, commit, and retrieval date.
• Positioning table distinguishes user segment from technical implementation.
• Strengths are tied to user-visible behavior or code evidence.
• Weaknesses/gaps cite evidence or are labeled as 待验证.
• Opportunities cite the evidence rows they derive from.
• Risks and assumptions include the next verification step.

Common Fixes

Unsupported Architecture Claim

Before:

## Architecture
The app probably uses a microservice architecture.

After:

## Architecture
The repository exposes one Vite app and one Node server entrypoint:
- `apps/web/package.json:7` defines `vite --host`.
- `server/index.ts:1-42` creates the HTTP server.

Unsourced Comparison Row

Before:

| Dimension | Competitor | Our product |
|---|---|---|
| Export | HTML and PDF | HTML and PDF |

After:

| Dimension | Competitor | Source | Our product | Source |
|---|---|---|---|---|
| Export | HTML and PDF | `src/export.ts:12-84` | HTML and PDF | `src/export/share.ts:1-92` |

Stale Local Clone

Before:

Analyzed local copy in ~/Downloads/repo.

After:

Analyzed `$COMPETITORS_BASE/{product}/{owner-repo}` at commit
`<full-hash>` (`git log -1 --format='%H %cI %s'`).