跳到正文
cccskills
免费在 GitHub 公开

prepare-walkthrough-presentation

Transforms validated /linear-walkthrough artifacts into one presentation-ready deck outline per major codebase component. Use when the user asks to create a presentation, prepare a walkthrough deck, build an onboarding deck, produce an architecture review deck, or summarize a codebase for a technical audience. Requires walkthrough directory output -- reads unified walkthrough, per-section files, validation reports, coverage maps, entry points, and open questions -- then orchestrates four parallel agent phases to produce slide outlines with speaker notes, evidence references, and suggested visuals.

中文简介正在准备中。当前显示原文说明。

查看安装方法

包含的文件(3)

  • SKILL.md9.1 KB
  • references/agent-instructions.md9.5 KB
  • references/output-format.md6.9 KB

SKILL.md(原文)

安装前可以查看交给 AI 代理的指令内容。

Prepare Walkthrough Presentation

Transform validated /linear-walkthrough output into one presentation-ready deck outline per major codebase component. Each deck explains what the component is, how it works, how it connects to the system, how it is developed and operated, and what risks remain.

Walkthrough directory: <walkthrough-directory> (default: ./walkthrough/ in the current directory).

Output directory: <walkthrough-directory>/presentation/ (created if it does not exist).

Workflow

flowchart TD
    Start(["Invoke /prepare-walkthrough-presentation"]) --> P1
    P1["Phase 1: Component Extraction and Deck Planning<br>1 agent — read walkthrough outputs, identify components,<br>build deck plan per component"] --> P1Out
    P1Out["Artifacts:<br>presentation/component-index.md<br>presentation/deck-plans/component-deck-plan-{name}.md"] --> P2
    P2["Phase 2: Deck Generation<br>N parallel agents — build narrative and generate<br>full deck outline per component in one pass"] --> P2Out
    P2Out["Artifacts:<br>presentation/decks/component-deck-outline-{name}.md"] --> P3
    P3["Phase 3: Validation<br>M parallel agents — fact-check decks against<br>walkthrough sources, check cross-deck consistency"] --> P3Out
    P3Out["Artifacts:<br>presentation/validation/component-deck-validation-{name}.md"] --> P4
    P4["Phase 4: Packaging<br>1 agent — produce crosswalk index, apply corrections,<br>finalize all deck outlines"] --> P4Out
    P4Out["Artifacts:<br>presentation/presentation-crosswalk.md<br>presentation/decks/ (finalized)"] --> Done(["Complete"])

Inputs

Read these artifacts from the walkthrough directory (all paths below are relative to that directory):

ArtifactPathRequired
Unified walkthroughunified-walkthrough.md or unified/index.mdYes
Walkthrough sectionssections/walkthrough-section-*.mdYes
Validation reportsvalidation/validation-report-*.mdYes
Coverage plancoverage-plan.mdYes
Entry pointsentry-points.mdYes
Open questionsopen-questions.mdYes

Input Validation

Before starting Phase 1, verify all required walkthrough artifacts exist. Check for unified-walkthrough.md (or unified/index.md), at least one file matching sections/walkthrough-section-*.md, coverage-plan.md, entry-points.md, and open-questions.md. If any required artifact is missing, stop and report which files are absent.

Phase 1: Component Extraction and Deck Planning

Spawn one general-purpose agent to read walkthrough outputs and produce a component index and per-component deck plans.

Agent prompt context

Include these in the agent prompt (pass file paths — do not transcribe file contents):

  • Walkthrough directory path
  • File path: agent-instructions.md — agent reads section "Planning Agent Instructions"
  • File path: output-format.md — agent reads sections "Component Index Format" and "Deck Plan Format"

Agent deliverables

ArtifactPath
Component indexpresentation/component-index.md
Deck planspresentation/deck-plans/component-deck-plan-{name}.md (one per component)

Orchestrator actions after Phase 1

  1. Read presentation/component-index.md to get the component list and deck-to-component mapping.
  2. Proceed to Phase 2 with one agent per component. Each agent reads its own deck plan.

Phase 2: Deck Generation

Spawn N parallel general-purpose agents — one per component from the deck plan. Each agent constructs the presentation narrative and generates the full deck outline in a single pass, eliminating the intermediate narrative artifact.

Agent prompt context (per agent)

Include these in the agent prompt (pass file paths — do not transcribe file contents):

  • Component deck plan file path
  • Walkthrough directory path (for reading source sections and evidence references)
  • File path: agent-instructions.md — agent reads section "Deck Generation Agent Instructions"
  • File path: output-format.md — agent reads sections "Deck Outline Format" and "Slide Format"

Agent deliverables (per agent)

ArtifactPath
Deck outlinepresentation/decks/component-deck-outline-{name}.md

Orchestrator actions after Phase 2

  1. Verify all expected deck outline files exist.
  2. Proceed to Phase 3.

Phase 3: Validation

Spawn M parallel general-purpose agents. Each validator checks one or more deck outlines against the source walkthrough.

Validator cross-assignment: rotate deck assignments so validator 1 checks decks from generation agent 2, validator 2 checks decks from generation agent 3, and so on (wrapping around). If there are fewer validators than generation agents, each validator checks multiple decks. If there is only one deck, the single validator checks it (self-validation is acceptable when no alternative exists).

Agent prompt context (per validator)

Include these in the agent prompt (pass file paths — do not transcribe file contents):

  • Assigned deck outline file paths
  • Walkthrough directory path
  • Other deck outline file paths (for cross-deck consistency checks)
  • File path: agent-instructions.md — agent reads section "Deck Validation Agent Instructions"
  • File path: output-format.md — agent reads section "Deck Validation Report Format"

Agent deliverables (per validator)

ArtifactPath
Validation reportpresentation/validation/component-deck-validation-{name}.md

Orchestrator actions after Phase 3

  1. Read all validation reports.
  2. If critical corrections exist (unsupported claims, invented architecture, incorrect sequencing), spawn a new general-purpose agent per affected deck to apply the corrections. Pass the agent the validation report and deck outline file path. The agent edits the deck file in place.
  3. Proceed to Phase 4.

Phase 4: Packaging

Spawn one general-purpose agent to finalize all deck outlines and produce the crosswalk index.

Agent prompt context

Include these in the agent prompt (pass file paths — do not transcribe file contents):

  • Directory path: presentation/decks/ — agent reads all deck outline files
  • Directory path: presentation/validation/ — agent reads all validation reports
  • File path: presentation/component-index.md
  • File path: agent-instructions.md — agent reads section "Packaging Agent Instructions"
  • File path: output-format.md — agent reads section "Presentation Crosswalk Format"

Agent deliverables

ArtifactPath
Presentation crosswalkpresentation/presentation-crosswalk.md
Finalized deck outlinespresentation/decks/ (corrections applied)

Large output handling

If the crosswalk or any deck outline exceeds 25k characters, apply the large file write strategy using Strategy A (multi-file split).

Deck Planning Heuristics

  • One deck per major component by default.
  • Merge when two components are too small or too tightly coupled to explain separately.
  • Split when one component is too broad to explain coherently in one deck.
  • A component may be an app, service, package, library, worker, pipeline, platform layer, or other meaningful architectural unit.

Default Audience

Technical audience: engineers, tech leads, platform owners, SRE/DevOps engineers, and engineering managers.

Presentation Style

  • Optimize for technical clarity, not executive fluff.
  • Use concise, high-signal slide text — avoid dense paragraphs.
  • Put detailed explanation in speaker notes.
  • Make each deck readable standalone but consistent with other decks.
  • Use consistent naming, terms, and structural patterns across all decks.

Evidence Handling

  • Every substantive slide references the walkthrough sections, validation reports, or source artifacts that support it.
  • Distinguish verified facts from inferred connections.
  • Preserve unresolved uncertainty from the walkthrough — do not smooth it away.

Resources

  • Agent instructions — detailed prompts for each agent type (planning, deck generation, validation, packaging)
  • Output format — required structure and templates for all artifacts

评价

还没有评价。欢迎分享你的使用感受。

评论

无需登录即可发表。欢迎分享使用感受、提问,或说明在哪些环境下无法运行。

还没有评论。来写第一条吧。

提交后会立即公开。请勿填写邮箱、电话等个人信息。URL 最多 2 个。

0 / 1000 字

此表单受 reCAPTCHA 保护,适用 Google 的隐私权政策和服务条款。

同一仓库的技能

Add automated documentation updater to any Claude skill. Creates a Python sync script that downloads upstream docs, processes markdown for AI consumption, and maintains local cache with configurable refresh. Collects template variables, then delegates implementation through 5-phase workflow. Use when adding auto-updating reference documentation to plugins or skills.

中文简介正在准备中。当前显示原文说明。

Jamie-BitFlight/claude_skills672026年10月20日 更新

SAM-style feature initiation workflow — discovery through codebase analysis, architecture spec, task decomposition, validation, and context manifest. Use when a user asks to add a feature, plan a feature, or convert an idea into an executable SAM plan.

中文简介正在准备中。当前显示原文说明。

Jamie-BitFlight/claude_skills672026年10月20日 更新

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include requests to "open a website", "fill out a form", "click a button", "take a screenshot", "scrape data from a page", "test this web app", "login to a site", "automate browser actions", or any task requiring programmatic web interaction.

中文简介正在准备中。当前显示原文说明。

Jamie-BitFlight/claude_skills672026年10月20日 更新

Browser automation for AI agents using the agent-browser CLI and Playwright. Use when navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, logging into sites, or automating any browser task. Triggers on "open a website", "fill out a form", "click a button", "scrape data", "test this web app", "automate browser actions", or any programmatic web interaction request.

中文简介正在准备中。当前显示原文说明。

Jamie-BitFlight/claude_skills672026年10月20日 更新

Runs the description-drift experiment — spawns all Claude Code agents simultaneously to collect self-reported capabilities, then compares them against static frontmatter descriptions to reveal how reliable orchestrator routing based on descriptions actually is. Use when measuring description drift across the agent fleet, re-running the capability collection experiment, analyzing a specific agent's self-reported capabilities, or auditing whether frontmatter descriptions accurately reflect agent behavior.

中文简介正在准备中。当前显示原文说明。

Jamie-BitFlight/claude_skills672026年10月20日 更新

Create or adapt Claude Code agent definitions. Use when creating an agent, changing subagent configuration, selecting agent scope, or designing a specialized delegation role.

中文简介正在准备中。当前显示原文说明。

Jamie-BitFlight/claude_skills672026年10月20日 更新

查看 Jamie-BitFlight 的全部技能

报告此技能的问题