Home / All Skills / Docs & Knowledge / competitive-teardown
Docs & Knowledge · nexu-io/html-anything

competitive-teardown

Positioning map + feature matrix + pricing comparison + opportunity windows: transforms competitor materials into an actionable product decision report.

L1 Basic · Source registration onlyV1 Commit-PinnedDual-pass Audit
Author nexu-ioGitHub nexu-io/html-anything ↗Stars 8827License Apache-2.0(仓库根 LICENSE:Apache License Version 2.0, January 2004;GitHub API spdx_id=Apache-2.0)Commit 553ed98c28
Control levels describe skill capabilities and suggested control measures; they do not represent quality or safety certifications, nor do they dictate usability. Usage decisions rest entirely with users and administrators. Different agent hosts apply varying constraints; do not assume hosts provide automatic protection. Third-party skill copyrights belong to their original authors. Use at your own risk.

1How It Works & Mechanism

The template itself contains no executable code: loader reads frontmatter and body from src/lib/templates/skills/<id>/SKILL.md, and returns body along with example content as LoadedSkill.

next/src/lib/templates/loader.ts
const raw = safeRead(path.join(dir, "SKILL.md"));
What this does: SKILL.md is the sole source of behavior—frontmatter is used for selector rendering (zh_name/emoji/aspect_hint/tags/example_*, etc.), while the entire body serves as the prompt; loader performs no semantic validation and dispatches whatever is written in the template.

The host assembles the prompt in a fixed manner before dispatch: the template body is sandwiched between SHARED_DESIGN_DIRECTIVES and [Input Format]/[User Content].

next/src/lib/templates/shared.ts
Wrap a per-template instruction body with the shared design directives and
What this does: The template only defines 'what this specific deliverable looks like' (layout/rhythm/components), while shared directives enforce cross-template hard constraints—single-file HTML, prohibition of file tools, Tailwind and Google Fonts loaded via CDN, content authenticity, and CJK spacing rules.

Shared directives formulate hard constraints as explicit prohibitions for the agent and enforce a content-driven quantity rule (numbers in templates serve only as reference lower bounds).

next/src/lib/templates/shared.ts
**禁止使用 Write / Edit / MultiEdit / Bash / Create / 任何文件系统工具**。
Later in the same file is another rule: the number of output slides/frames/cards/sections is entirely determined by the actual length and information structure of [User Content]—explaining why templates of this type do not break as content expands: at least one card/section per content chunk, with the template merely providing a layout pool.

Actual execution is performed by the coding agent CLI already installed and logged in on the user's local machine: the host spawns a child process, hands the prompt to it, and streams HTML output from stdout.

next/src/lib/agents/invoke.ts
import { spawn, type ChildProcessWithoutNullStreams } from "node:child_process";
What this does: The template skill only provides prompts; execution capabilities (including code writing and tool execution) come from the host CLI's existing permissions; the skill directory itself contains no script entry points, so grading evaluates only assets within the directory (host behaviors are excluded).

/api/convert is the execution path for templates: it assembles skill.body and user content into a prompt and passes it to invokeAgent.

next/src/app/api/convert/route.ts
prompt = assemblePrompt({ body: skill.body, content, format });
Edit mode uses buildEditPrompt (minimal-diff editing, file tools disabled, requiring direct streaming of the complete HTML), yielding a stable artifact profile: single-file HTML, inline CSS/SVG, and external links restricted solely to fonts and chart libraries explicitly requested by specific templates.

The core of the template is a fixed 9-section structure: from Header, Executive takeaway, Positioning map, all the way to 30/90/180-day Recommended moves; missing sections constitute failure.

next/src/lib/templates/skills/competitive-teardown/SKILL.md
【必须输出的结构】
What this does: Transforms competitive analysis from open-ended writing into a fill-in-the-blank report, where every section has strict content requirements (e.g., exactly 3 Executive takeaways, each must include a 'so what').

The positioning map and feature matrix are mandated to be generated from user content rather than template boilerplate: axes must derive from source materials, matrix columns include Us / Opportunity, and ✓/△/— are used for coverage.

next/src/lib/templates/skills/competitive-teardown/SKILL.md
3. Positioning map: 用 2×2 象限或坐标图表现竞品定位。坐标轴必须来自用户内容, 不要套模板词。
Rule 5 of the same section mandates matrix rows as key capabilities and columns as Competitors + Us / Opportunity, annotated with short notes—the matrix is a decision table, not a mere checklist.

One card per competitor with fixed fields: target user / core promise / pricing signal / primary strength / visible weakness; opportunity windows strictly include why now, target segment, first move, and risk.

next/src/lib/templates/skills/competitive-teardown/SKILL.md
8. Opportunity windows: 3 个机会窗口, 每个包含 why now、target segment、first move、risk。
Fixed fields ensure comparability across reports: teardowns across different runs can be read horizontally in the same structure; '3 opportunity windows' serves as a lower bound, paired with 30/90/180-day recommended actions.

Includes a dedicated content authenticity section: use only user-provided materials, label missing information as 'not found in source' / 'unknown', prohibit fabricating market share/ARR/customer names/pricing, and list gaps under Evidence gaps if data is insufficient.

next/src/lib/templates/skills/competitive-teardown/SKILL.md
- 只使用用户提供的竞品、价格、功能、评论。缺失信息用 "not found in source" 或 "unknown" 标注。
This is the most critical discipline of this template: turning 'no hallucinated figures' into an auditable output convention (manifesting as 'unknown' / 'not found' directly in the matrix), rather than a vague anti-hallucination request.

Three stylistic choices are provided via HTML references in assets/ (war-room-grid as default, radar-map dark intelligence console, analyst-dossier paper research report), requiring selection based on input characteristics without mixing.

next/src/lib/templates/skills/competitive-teardown/SKILL.md
- `assets/war-room-grid.html`: 默认风格。浅色战情室 / 咨询报告, 适合产品团队、PM、普通商业读者。
What this does: Concretizes a strategic consulting aesthetic into three pre-implemented style boilerplates (44-82 lines of HTML each), allowing the agent to copy variables and components directly rather than inventing them from scratch; this is also the only template in this batch with internal assets (local files, read internally without outbound transmission).

2Core Capabilities

01Fixed 9-section report structure (header / takeaway / positioning / cards / matrix / pricing / UX / opportunities / moves)
02Executive takeaway: exactly 3 assessments, each including a 'so what'
03Feature matrix (rows = key capabilities, columns = competitors + Us/Opportunity, ✓/△/— coverage + short notes)
04Pricing/packaging breakdown section (tiers, trials, limitations, enterprise sales motions)
053 selectable visual styles (war-room-grid / radar-map / analyst-dossier), defaulting to war-room-grid
06Content authenticity discipline (label missing data as unknown / not found in source, prohibit inventing numbers, still produce report and list Evidence gaps when info is sparse)
07Single-competitor fallback path (outputs a deep dive when only one competitor is provided)

4Recommended Governance & Evidence

Suggested Controls for Safe Operation
For Agent Runtime
Can be auto-loaded
For Administrators
Register source and version
Evidence & Triggers (File Paths & Excerpts)
  • SKILL.md: Pure prompt directives and rule text, without script execution or external side effects
Audit Notes: . Grading is based solely on the skill's own directory (contains only SKILL.md + example.html + example.md + 3 style reference HTMLs in assets/, no script entry points [assets and example each contain one Tailwind CDN <script> reference], no built-in CLI, no credential reading, and no disk write instructions): no script entry points, no bundled CLI, no credential access, and no disk write directives; SKILL.md body does not request importing any external endpoints at runtime (three styles are supplied as local assets/ files; SKILL.md defines only layout and content structure). No outbound transmission occurs during runtime. Endpoints present in the directory exist exclusively in assets/*.html and example.html referencing cdn.tailwindcss.com and fonts.googleapis.com/gstatic (browser-side asset fetching for style references and artifact samples). Per Batch 4 final criteria, these constitute **artifact-time references** (fetched by the browser upon opening the generated HTML; the skill does not scrape, report, or transmit data during execution), hence they do not impact security grading and are only cataloged here; host wrapper mechanics (loader -> assemblePrompt -> local agent CLI spawn, preview iframe sandbox) are host behaviors and recorded in verdict.risks.

5Second-Pass Verification

  • [discrepancy] 九段结构在示例中的覆盖度 — example.html 展示了 Header、Executive takeaway(3 条)、Positioning map、3 张竞品卡、Feature matrix(含 not found / unknown 单元格)、30/90/180 天 moves;但第 6 段 Pricing / packaging read 与第 7 段 UX / messaging notes 未独立成节,第 8 段 Opportunity windows 亦未以独立小节出现(部分内容并入卡片与矩阵注释)。模板能力以 SKILL.md 为准,示例属精简演示。
  • [ok] 内容真实性纪律的可见落地 — 矩阵中逐字出现 unknown(Transcript library 行)与 not found(CRM / sales workflow、Cross-interview insight cards 行),与 SKILL.md 要求的标注词一致,证明该纪律可被执行。
  • [ok] 三套 assets 是否为真实不同风格 — war-room-grid:浅色 --paper:#f7f4ec + Inter;radar-map:深色 panel rgba(13,27,45,.76) + 径向雷达网格 repeating-radial-gradient;analyst-dossier:双线规则 .rule + 红色 stamp 旋转 -2deg + IBM Plex Serif 衬线。三份互不相同的实现坐实风格三选一。
  • [ok] 端点与判级 — SKILL.md 正文无 URL/端点要求;端点只出现在 assets 与 example 的 cdn.tailwindcss.com <script> 及字体链接,属产物期引用层,执行期零外发,维持 green。
  • [ok] license / stars / commit 元数据 — GitHub API(api.github.com/repos/nexu-io/html-anything):spdx_id=Apache-2.0、stargazers_count=8827、pushed_at=2026-09-15T03:36:24Z;工作副本 git HEAD = 553ed98c283f9c0f489902d035416a972d6a9699(与任务表 pin 553ed98c283f 同源)。
  • [ok] skill.path 与目录资产面 — pin 目录内该路径存在且含 SKILL.md;目录内文件清单已用 glob 全量核对(含隐藏文件),无 scripts/、无 references/、无二进制资产(competitive-teardown 的 assets/ 为纯 HTML,另行记录)。

6Verdict & Scenarios

  • Structure guarantees quality: the fixed 9-section skeleton (including quantity and field constraints) turns open-ended competitive analysis into a deliverable, meeting-ready report.
  • Anti-hallucination discipline is embedded in the template and auditable (unknown / not found in source, prohibition on inventing ARR/pricing/customer names, listing Evidence gaps when data is insufficient).
  • Design-side provides three pre-implemented style references (light war room / dark radar console / paper dossier) and mandates choosing one based on input, balancing visual consistency with stylistic flexibility.
  • Clear fallback path: optimal for 2–6 competitors, shifts to deep dive for a single competitor, avoiding delivery refusal due to limited input scope.
  • The matrix guides competitor data directly into strategic decisions via Us / Opportunity columns, preventing the report from lingering at a purely descriptive level.
  • Best for: Ideal for product and strategy teams turning fragmented competitor materials (official websites, pricing pages, changelogs, reviews, sales feedback) into an actionable teardown report; also suitable for preliminary structuring in equity research/industry analysis (using analyst-dossier style). Performs best with 2–6 competitors having pricing, feature, and review data.
    Not for: Not suitable for marketing copy or generic articles (explicitly excluded in SKILL.md); not suitable for scenarios lacking source materials where the model is expected to perform web research autonomously (the template strictly consumes user-provided data); not suitable for compliance reports requiring formal page citations (the template does not generate citation indices).
    Install Agent-Ready
    ① Verified Mirror 2026-09-15
    Option A · Direct DownloadDownload competitive-teardown.tar.gz
    sha256: 99cee63fe9181008…
    Option B · Copy Guide for Agent
    install.json
    Agents automatically download and verify mirrors from this JSON guide.
    ② Upstream GitHub
    GitHub Repo ↗
    Pinned to commit 553ed98c28.