[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M6NQ5VSH5M682YMYDF9H9KFE":3,"skill-same:01M6NQ5VSH5M682YMYDF9H9KFE":72},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":23,"publishedAt":44,"linkOnly":45,"markdownContent":46,"markdownBody":47,"frontMatter":48,"refName":51,"commitSha":52,"skillDirectory":53,"forksCount":14,"repositoryHtmlUrl":54,"repositoryHomepageUrl":6,"ownerHtmlUrl":55,"ownerBlogUrl":56,"ownerDescription":6,"updatedAt":57,"install":58,"resources":71},"01M6NQ5VSH5M682YMYDF9H9KFE","app-workshop-session-pdf-recap-next",null,"Extract content from a workshop session PDF and produce a structured recap focused on (1) what was learned, (2) next steps, and (3) possible enhancements. Ingests a single PDF file path, parses session material (slides, transcripts, exercises, Q&A), and emits a Markdown report with Learning Recap, Next Steps, and Enhancement Opportunities sections with page-level traceability. Use when the user asks to \"summarize a workshop\", \"recap a training PDF\", \"pull next steps from a session deck\", or provides a workshop\u002Fseminar\u002Ftutorial PDF and wants a learn-and-act summary.","carlosmarte\u002Fai-agent-artifacts","carlosmarte","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F3478576?v=4","01M6NPRTEXZG86XJXMGQ05A24X","Carlos Marte","01M6NPRTK4A6A4Q8CMWMH1ZY7W",0,"MIT","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fapp-workshop-session-pdf-recap-next\u002FSKILL.md",[18],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000002","文档处理","SCENARIO","使用场景",[24,25,26,27,28,29,30,31,32,33,34,35,36,37,38,39,40,41,42,43],"app workshop session pdf recap next","app","next","recap","session","PDF","tutorial","next steps","traceability","material","false","deck","use","page","slides","learning","produce","pull","learn","training","2026-11-10 10:12:30",false,"---\nname: app-workshop-session-pdf-recap-next\ndescription: Extract content from a workshop session PDF and produce a structured recap focused on (1) what was learned, (2) next steps, and (3) possible enhancements. Ingests a single PDF file path, parses session material (slides, transcripts, exercises, Q&A), and emits a Markdown report with Learning Recap, Next Steps, and Enhancement Opportunities sections with page-level traceability. Use when the user asks to \"summarize a workshop\", \"recap a training PDF\", \"pull next steps from a session deck\", or provides a workshop\u002Fseminar\u002Ftutorial PDF and wants a learn-and-act summary.\ntier: app\nallowed-tools: Read, Glob, Bash, Write\nargument-hint: \"\u003Cpdf_path> [output_path]\"\ndisable-model-invocation: false\n---\n\n# Workshop Session PDF — Recap & Next Steps\n\nExtract content from a single workshop session PDF and produce a focused recap organized around three questions:\n\n1. **What was learned?** — the durable takeaways, concepts, tools, and techniques surfaced in the session.\n2. **What are the next steps?** — the actions, follow-ups, and practice work the attendee should do after the session.\n3. **What enhancements could be added?** — improvements, extensions, or deeper explorations beyond the material presented.\n\nThis skill is **read-only** with respect to the input PDF. It produces a Markdown report and does not modify the source file.\n\n## Inputs\n\nArguments are parsed positionally:\n\n- `\u003Cpdf_path>` (required) — absolute path to a single `.pdf` file containing workshop session material (slides, handouts, transcript, lab guide, or any combination).\n- `[output_path]` (optional) — absolute path where the recap Markdown should be written. If omitted, derive a unique default so repeated runs never overwrite a prior recap:\n  - Path: `\u003Cpdf_path_dirname>\u002F\u003Cpdf_basename>-recap-\u003CYYYYMMDD-HHMMSS>-\u003CUUID8>.md`\n  - `YYYYMMDD-HHMMSS` is the local-time timestamp at the start of the run (generate once via `Bash` with `date +%Y%m%d-%H%M%S` and reuse).\n  - `UUID8` is an 8-hex-char random suffix (generate via `Bash` with `uuidgen | tr -d '-' | cut -c1-8 | tr A-Z a-z`).\n  - Announce the resolved output path to the user before writing.\n\nIf `\u003Cpdf_path>` is missing, not a file, not readable, or does not end in `.pdf`, stop and ask the user to supply a valid path. Do not attempt to infer the PDF from nearby files.\n\n## Execution Protocol\n\nRun the five phases in order. Each phase's output feeds the next — do not skip.\n\n### Phase 1 — PDF Ingestion\n\n1. Verify the PDF exists and is readable (`Bash` with `ls -l \u003Cpdf_path>` is sufficient).\n2. Determine the page count before reading. Use `Bash` with `mdls -name kMDItemNumberOfPages \u003Cpdf_path>` on macOS, or fall back to reading the PDF incrementally.\n3. Read the PDF content via the `Read` tool:\n   - For PDFs ≤ 10 pages: read without the `pages` parameter.\n   - For PDFs > 10 pages: read in ranges of up to 20 pages per call (e.g., `pages: \"1-20\"`, then `pages: \"21-40\"`, …) until all pages are consumed. The Read tool requires the `pages` parameter for PDFs longer than 10 pages.\n4. While reading, capture page-anchored notes in a working scratchpad:\n   - Session title, presenter(s), date, venue (if shown on cover\u002Ffooter).\n   - Agenda or table of contents (if present).\n   - Per-section topics, key terms, diagrams, and code\u002Flab references, each tagged with its page number(s).\n   - Explicit \"takeaways\", \"key points\", \"summary\", \"homework\", \"next steps\", \"further reading\" slides — these are primary source material for the recap and must be captured verbatim or near-verbatim.\n5. Do **not** fabricate content. If a detail is not visible in the PDF, mark it `NOT IN SOURCE` rather than inferring from training data or prior conversations.\n\n### Phase 2 — Session Frame\n\nProduce a short session frame block (emitted later as `## Session`) capturing:\n\n- **Title** — from the cover slide or document metadata; `NOT IN SOURCE` if absent.\n- **Presenter(s)** — as printed; `NOT IN SOURCE` if absent.\n- **Date \u002F Event** — as printed; `NOT IN SOURCE` if absent.\n- **Format** — e.g., slide deck, lab guide, workbook, transcript, hybrid (infer from the structure you observed, not the filename).\n- **Page count** — total pages read.\n- **Scope** — one sentence describing the stated subject of the session.\n\n### Phase 3 — Learning Recap Extraction\n\nUnder `## Learning Recap`, extract what was actually taught. Each entry must be traceable.\n\nGroup by:\n\n- **Concepts & Frameworks** — ideas, models, mental frameworks introduced.\n- **Tools & Techniques** — specific tools, commands, libraries, or step-by-step techniques demonstrated.\n- **Examples & Exercises** — worked examples, labs, demos, or exercises walked through.\n- **Quoted Takeaways** — any explicit \"key point\", \"remember this\", or summary bullets from the source, quoted near-verbatim.\n\nFor each entry emit:\n\n```\nL-00X — \u003Cone-line description of the learning>\n  Evidence: \u003Cpdf_basename>:p\u003Cpage>[, p\u003Cpage>…]\n  Category: [Concept | Tool | Example | Takeaway]\n```\n\nDo not include every bullet on every slide — select the items that a motivated attendee would transcribe into their own notes. Aim for a dense but curated list, not an exhaustive transcript.\n\n### Phase 4 — Next Steps Extraction\n\nUnder `## Next Steps`, capture the actions the session implies or prescribes for the attendee. Sources of next steps, in priority order:\n\n1. **Explicit** — slides literally titled \"Next Steps\", \"Homework\", \"Action Items\", \"Try This\", \"Further Reading\", \"Exercises to Do\". These are authoritative and must be captured.\n2. **Implicit** — actions the material strongly implies (e.g., a tool was introduced but not practiced → \"practice with \u003Ctool>\"; a concept was taught with a promised follow-up → \"review \u003Ctopic>\").\n3. **Referenced resources** — links, books, docs, repos, or papers cited in the deck that the attendee should consult.\n\nFor each entry emit:\n\n```\nN-00X — \u003Cimperative action starting with a verb>\n  Source: [Explicit | Implicit | Referenced]\n  Evidence: \u003Cpdf_basename>:p\u003Cpage> (or \"derived from p\u003Cpage> — \u003Creason>\" for Implicit)\n  Effort: [Quick (\u003C1h) | Medium (half-day) | Deep (multi-day)]\n```\n\nIf the session prescribed zero explicit next steps and you cannot defensibly derive any implicit ones, say so — do not manufacture tasks.\n\n### Phase 5 — Enhancement Opportunities\n\nUnder `## Enhancement Opportunities`, propose extensions **beyond** the material. These are your recommendations, not content from the PDF. Each proposal must be anchored to something the session did present — this prevents drift into unrelated territory.\n\nCategorize each as:\n\n- **Depth** — go deeper on a topic the session introduced but did not fully develop.\n- **Breadth** — explore an adjacent topic the session gestured at but did not cover.\n- **Applied** — apply the learning to a realistic project or scenario not used in the session.\n- **Tooling** — automate, script, or templatize a manual process demonstrated in the session.\n\nFor each entry emit:\n\n```\nE-00X — \u003Cproposed enhancement>\n  Category: [Depth | Breadth | Applied | Tooling]\n  Anchor: \u003Cwhich learning this builds on — reference an L-00X ID from Phase 3>\n  Rationale: \u003Cone sentence: why this adds value beyond the session>\n```\n\nIf an enhancement cannot be anchored to an L-00X ID, drop it — unanchored suggestions are out of scope.\n\n## Output Contract\n\nWrite the recap to `\u003Coutput_path>` (or the default derived path) as a single Markdown document with these top-level sections, in this order:\n\n1. `## Session` — frame block from Phase 2.\n2. `## Learning Recap` — grouped L-00X entries from Phase 3.\n3. `## Next Steps` — N-00X entries from Phase 4, ordered: all Explicit first, then Implicit, then Referenced.\n4. `## Enhancement Opportunities` — E-00X entries from Phase 5, grouped by category.\n5. `## Traceability Index` — a table mapping every L \u002F N \u002F E ID to its page reference (or anchor ID for enhancements) and its category, so each row is verifiable against the source PDF.\n\nAfter writing the file, print a one-paragraph summary to the conversation: the output file path, the counts of L \u002F N \u002F E entries, and one sentence naming the session's central theme. Do not repeat the full document inline.\n\n## Operational Boundaries\n\n- **Do not modify the source PDF** or any file in its directory other than the output recap.\n- **Do not read files other than the specified PDF.** No sibling scanning, no prior-session comparisons, no repo exploration. The input is one PDF; the output is one Markdown file.\n- **Do not invent content.** Every L-00X and N-00X (Explicit\u002FReferenced) entry must cite a page. Implicit next steps must name the page that implied them. Enhancements must anchor to an L-00X ID.\n- **Do not expand scope beyond recap + next steps + enhancements.** Resist adding glossaries, cheat sheets, quizzes, or presenter critiques unless the user explicitly asks.\n- **Stop and ask** when: the PDF is unreadable, encrypted, empty, image-only with no extractable text (OCR is out of scope), or clearly not workshop material (e.g., a legal contract, invoice, novel).\n\n## Example Invocation\n\n```\n\u002Fworkshop-session-pdf-recap-next \u002FUsers\u002Fme\u002FDownloads\u002Fclaude-agents-workshop.pdf\n```\n\nExpected output skeleton (written to, e.g., `\u002FUsers\u002Fme\u002FDownloads\u002Fclaude-agents-workshop-recap-20260421-143522-a1b2c3d4.md`):\n\n```\n## Session\n- Title: Building Claude Agents — Hands-On Workshop\n- Presenter(s): J. Doe, A. Smith\n- Date \u002F Event: 2026-03-14, Anthropic DevDay\n- Format: Slide deck + lab guide (hybrid)\n- Page count: 42\n- Scope: Introduces the Agent SDK and walks attendees through building a research agent end-to-end.\n\n## Learning Recap\n\n### Concepts & Frameworks\nL-001 — Agent loop: plan → act → observe → reflect\n  Evidence: claude-agents-workshop.pdf:p4, p5\n  Category: Concept\n\nL-002 — Tool-use contract: schema-first tool definitions\n  Evidence: claude-agents-workshop.pdf:p11\n  Category: Concept\n\n### Tools & Techniques\nL-003 — Using `ToolSearch` to defer schema loading until needed\n  Evidence: claude-agents-workshop.pdf:p17, p18\n  Category: Tool\n\n### Examples & Exercises\nL-004 — Lab: build a 3-tool research agent that summarizes a URL\n  Evidence: claude-agents-workshop.pdf:p22-p27\n  Category: Example\n\n### Quoted Takeaways\nL-005 — \"Prefer small, composable tools over one monolithic tool.\" (slide 31)\n  Evidence: claude-agents-workshop.pdf:p31\n  Category: Takeaway\n\n## Next Steps\nN-001 — Complete the post-workshop lab in the provided repo\n  Source: Explicit\n  Evidence: claude-agents-workshop.pdf:p40\n  Effort: Medium (half-day)\n\nN-002 — Read the Agent SDK reference docs sections listed on slide 41\n  Source: Referenced\n  Evidence: claude-agents-workshop.pdf:p41\n  Effort: Quick (\u003C1h)\n\nN-003 — Practice defining a new tool schema from scratch\n  Source: Implicit\n  Evidence: derived from p11 — tool-use contract was introduced but not rehearsed by the attendee\n  Effort: Quick (\u003C1h)\n\n## Enhancement Opportunities\n\n### Depth\nE-001 — Add structured output (JSON schema) to the research agent from the lab\n  Category: Depth\n  Anchor: L-004\n  Rationale: The lab stops at free-text output; JSON output unlocks downstream automation.\n\n### Applied\nE-002 — Adapt the agent loop to a PR-review use case against a real repo\n  Category: Applied\n  Anchor: L-001\n  Rationale: The session used a research example; porting the loop to code review tests generalization.\n\n### Tooling\nE-003 — Package the 3-tool stack as a reusable npm module with tests\n  Category: Tooling\n  Anchor: L-003\n  Rationale: Turns one-off lab code into a durable library the attendee can reuse.\n\n## Traceability Index\n| ID    | Type        | Category \u002F Source  | Evidence                                |\n|-------|-------------|--------------------|-----------------------------------------|\n| L-001 | Learning    | Concept            | claude-agents-workshop.pdf:p4, p5       |\n| L-002 | Learning    | Concept            | claude-agents-workshop.pdf:p11          |\n| L-003 | Learning    | Tool               | claude-agents-workshop.pdf:p17, p18     |\n| L-004 | Learning    | Example            | claude-agents-workshop.pdf:p22-p27      |\n| L-005 | Learning    | Takeaway           | claude-agents-workshop.pdf:p31          |\n| N-001 | Next Step   | Explicit           | claude-agents-workshop.pdf:p40          |\n| N-002 | Next Step   | Referenced         | claude-agents-workshop.pdf:p41          |\n| N-003 | Next Step   | Implicit           | derived from p11                        |\n| E-001 | Enhancement | Depth → L-004      | anchor only                             |\n| E-002 | Enhancement | Applied → L-001    | anchor only                             |\n| E-003 | Enhancement | Tooling → L-003    | anchor only                             |\n```\n\n## Notes on PDF Handling\n\n- The `Read` tool natively supports PDF. For PDFs > 10 pages, the `pages` parameter is **required** (max 20 pages per call). Plan your reads before starting so you cover every page exactly once.\n- If the PDF is image-only (scanned slides with no text layer), the `Read` tool will return little or no text. In that case, stop and report — OCR is out of scope for this skill.\n- Treat the PDF basename (without extension) as the default slug when constructing the output path and when labelling evidence.\n","# Workshop Session PDF — Recap & Next Steps\n\nExtract content from a single workshop session PDF and produce a focused recap organized around three questions:\n\n1. **What was learned?** — the durable takeaways, concepts, tools, and techniques surfaced in the session.\n2. **What are the next steps?** — the actions, follow-ups, and practice work the attendee should do after the session.\n3. **What enhancements could be added?** — improvements, extensions, or deeper explorations beyond the material presented.\n\nThis skill is **read-only** with respect to the input PDF. It produces a Markdown report and does not modify the source file.\n\n## Inputs\n\nArguments are parsed positionally:\n\n- `\u003Cpdf_path>` (required) — absolute path to a single `.pdf` file containing workshop session material (slides, handouts, transcript, lab guide, or any combination).\n- `[output_path]` (optional) — absolute path where the recap Markdown should be written. If omitted, derive a unique default so repeated runs never overwrite a prior recap:\n  - Path: `\u003Cpdf_path_dirname>\u002F\u003Cpdf_basename>-recap-\u003CYYYYMMDD-HHMMSS>-\u003CUUID8>.md`\n  - `YYYYMMDD-HHMMSS` is the local-time timestamp at the start of the run (generate once via `Bash` with `date +%Y%m%d-%H%M%S` and reuse).\n  - `UUID8` is an 8-hex-char random suffix (generate via `Bash` with `uuidgen | tr -d '-' | cut -c1-8 | tr A-Z a-z`).\n  - Announce the resolved output path to the user before writing.\n\nIf `\u003Cpdf_path>` is missing, not a file, not readable, or does not end in `.pdf`, stop and ask the user to supply a valid path. Do not attempt to infer the PDF from nearby files.\n\n## Execution Protocol\n\nRun the five phases in order. Each phase's output feeds the next — do not skip.\n\n### Phase 1 — PDF Ingestion\n\n1. Verify the PDF exists and is readable (`Bash` with `ls -l \u003Cpdf_path>` is sufficient).\n2. Determine the page count before reading. Use `Bash` with `mdls -name kMDItemNumberOfPages \u003Cpdf_path>` on macOS, or fall back to reading the PDF incrementally.\n3. Read the PDF content via the `Read` tool:\n   - For PDFs ≤ 10 pages: read without the `pages` parameter.\n   - For PDFs > 10 pages: read in ranges of up to 20 pages per call (e.g., `pages: \"1-20\"`, then `pages: \"21-40\"`, …) until all pages are consumed. The Read tool requires the `pages` parameter for PDFs longer than 10 pages.\n4. While reading, capture page-anchored notes in a working scratchpad:\n   - Session title, presenter(s), date, venue (if shown on cover\u002Ffooter).\n   - Agenda or table of contents (if present).\n   - Per-section topics, key terms, diagrams, and code\u002Flab references, each tagged with its page number(s).\n   - Explicit \"takeaways\", \"key points\", \"summary\", \"homework\", \"next steps\", \"further reading\" slides — these are primary source material for the recap and must be captured verbatim or near-verbatim.\n5. Do **not** fabricate content. If a detail is not visible in the PDF, mark it `NOT IN SOURCE` rather than inferring from training data or prior conversations.\n\n### Phase 2 — Session Frame\n\nProduce a short session frame block (emitted later as `## Session`) capturing:\n\n- **Title** — from the cover slide or document metadata; `NOT IN SOURCE` if absent.\n- **Presenter(s)** — as printed; `NOT IN SOURCE` if absent.\n- **Date \u002F Event** — as printed; `NOT IN SOURCE` if absent.\n- **Format** — e.g., slide deck, lab guide, workbook, transcript, hybrid (infer from the structure you observed, not the filename).\n- **Page count** — total pages read.\n- **Scope** — one sentence describing the stated subject of the session.\n\n### Phase 3 — Learning Recap Extraction\n\nUnder `## Learning Recap`, extract what was actually taught. Each entry must be traceable.\n\nGroup by:\n\n- **Concepts & Frameworks** — ideas, models, mental frameworks introduced.\n- **Tools & Techniques** — specific tools, commands, libraries, or step-by-step techniques demonstrated.\n- **Examples & Exercises** — worked examples, labs, demos, or exercises walked through.\n- **Quoted Takeaways** — any explicit \"key point\", \"remember this\", or summary bullets from the source, quoted near-verbatim.\n\nFor each entry emit:\n\n```\nL-00X — \u003Cone-line description of the learning>\n  Evidence: \u003Cpdf_basename>:p\u003Cpage>[, p\u003Cpage>…]\n  Category: [Concept | Tool | Example | Takeaway]\n```\n\nDo not include every bullet on every slide — select the items that a motivated attendee would transcribe into their own notes. Aim for a dense but curated list, not an exhaustive transcript.\n\n### Phase 4 — Next Steps Extraction\n\nUnder `## Next Steps`, capture the actions the session implies or prescribes for the attendee. Sources of next steps, in priority order:\n\n1. **Explicit** — slides literally titled \"Next Steps\", \"Homework\", \"Action Items\", \"Try This\", \"Further Reading\", \"Exercises to Do\". These are authoritative and must be captured.\n2. **Implicit** — actions the material strongly implies (e.g., a tool was introduced but not practiced → \"practice with \u003Ctool>\"; a concept was taught with a promised follow-up → \"review \u003Ctopic>\").\n3. **Referenced resources** — links, books, docs, repos, or papers cited in the deck that the attendee should consult.\n\nFor each entry emit:\n\n```\nN-00X — \u003Cimperative action starting with a verb>\n  Source: [Explicit | Implicit | Referenced]\n  Evidence: \u003Cpdf_basename>:p\u003Cpage> (or \"derived from p\u003Cpage> — \u003Creason>\" for Implicit)\n  Effort: [Quick (\u003C1h) | Medium (half-day) | Deep (multi-day)]\n```\n\nIf the session prescribed zero explicit next steps and you cannot defensibly derive any implicit ones, say so — do not manufacture tasks.\n\n### Phase 5 — Enhancement Opportunities\n\nUnder `## Enhancement Opportunities`, propose extensions **beyond** the material. These are your recommendations, not content from the PDF. Each proposal must be anchored to something the session did present — this prevents drift into unrelated territory.\n\nCategorize each as:\n\n- **Depth** — go deeper on a topic the session introduced but did not fully develop.\n- **Breadth** — explore an adjacent topic the session gestured at but did not cover.\n- **Applied** — apply the learning to a realistic project or scenario not used in the session.\n- **Tooling** — automate, script, or templatize a manual process demonstrated in the session.\n\nFor each entry emit:\n\n```\nE-00X — \u003Cproposed enhancement>\n  Category: [Depth | Breadth | Applied | Tooling]\n  Anchor: \u003Cwhich learning this builds on — reference an L-00X ID from Phase 3>\n  Rationale: \u003Cone sentence: why this adds value beyond the session>\n```\n\nIf an enhancement cannot be anchored to an L-00X ID, drop it — unanchored suggestions are out of scope.\n\n## Output Contract\n\nWrite the recap to `\u003Coutput_path>` (or the default derived path) as a single Markdown document with these top-level sections, in this order:\n\n1. `## Session` — frame block from Phase 2.\n2. `## Learning Recap` — grouped L-00X entries from Phase 3.\n3. `## Next Steps` — N-00X entries from Phase 4, ordered: all Explicit first, then Implicit, then Referenced.\n4. `## Enhancement Opportunities` — E-00X entries from Phase 5, grouped by category.\n5. `## Traceability Index` — a table mapping every L \u002F N \u002F E ID to its page reference (or anchor ID for enhancements) and its category, so each row is verifiable against the source PDF.\n\nAfter writing the file, print a one-paragraph summary to the conversation: the output file path, the counts of L \u002F N \u002F E entries, and one sentence naming the session's central theme. Do not repeat the full document inline.\n\n## Operational Boundaries\n\n- **Do not modify the source PDF** or any file in its directory other than the output recap.\n- **Do not read files other than the specified PDF.** No sibling scanning, no prior-session comparisons, no repo exploration. The input is one PDF; the output is one Markdown file.\n- **Do not invent content.** Every L-00X and N-00X (Explicit\u002FReferenced) entry must cite a page. Implicit next steps must name the page that implied them. Enhancements must anchor to an L-00X ID.\n- **Do not expand scope beyond recap + next steps + enhancements.** Resist adding glossaries, cheat sheets, quizzes, or presenter critiques unless the user explicitly asks.\n- **Stop and ask** when: the PDF is unreadable, encrypted, empty, image-only with no extractable text (OCR is out of scope), or clearly not workshop material (e.g., a legal contract, invoice, novel).\n\n## Example Invocation\n\n```\n\u002Fworkshop-session-pdf-recap-next \u002FUsers\u002Fme\u002FDownloads\u002Fclaude-agents-workshop.pdf\n```\n\nExpected output skeleton (written to, e.g., `\u002FUsers\u002Fme\u002FDownloads\u002Fclaude-agents-workshop-recap-20260421-143522-a1b2c3d4.md`):\n\n```\n## Session\n- Title: Building Claude Agents — Hands-On Workshop\n- Presenter(s): J. Doe, A. Smith\n- Date \u002F Event: 2026-03-14, Anthropic DevDay\n- Format: Slide deck + lab guide (hybrid)\n- Page count: 42\n- Scope: Introduces the Agent SDK and walks attendees through building a research agent end-to-end.\n\n## Learning Recap\n\n### Concepts & Frameworks\nL-001 — Agent loop: plan → act → observe → reflect\n  Evidence: claude-agents-workshop.pdf:p4, p5\n  Category: Concept\n\nL-002 — Tool-use contract: schema-first tool definitions\n  Evidence: claude-agents-workshop.pdf:p11\n  Category: Concept\n\n### Tools & Techniques\nL-003 — Using `ToolSearch` to defer schema loading until needed\n  Evidence: claude-agents-workshop.pdf:p17, p18\n  Category: Tool\n\n### Examples & Exercises\nL-004 — Lab: build a 3-tool research agent that summarizes a URL\n  Evidence: claude-agents-workshop.pdf:p22-p27\n  Category: Example\n\n### Quoted Takeaways\nL-005 — \"Prefer small, composable tools over one monolithic tool.\" (slide 31)\n  Evidence: claude-agents-workshop.pdf:p31\n  Category: Takeaway\n\n## Next Steps\nN-001 — Complete the post-workshop lab in the provided repo\n  Source: Explicit\n  Evidence: claude-agents-workshop.pdf:p40\n  Effort: Medium (half-day)\n\nN-002 — Read the Agent SDK reference docs sections listed on slide 41\n  Source: Referenced\n  Evidence: claude-agents-workshop.pdf:p41\n  Effort: Quick (\u003C1h)\n\nN-003 — Practice defining a new tool schema from scratch\n  Source: Implicit\n  Evidence: derived from p11 — tool-use contract was introduced but not rehearsed by the attendee\n  Effort: Quick (\u003C1h)\n\n## Enhancement Opportunities\n\n### Depth\nE-001 — Add structured output (JSON schema) to the research agent from the lab\n  Category: Depth\n  Anchor: L-004\n  Rationale: The lab stops at free-text output; JSON output unlocks downstream automation.\n\n### Applied\nE-002 — Adapt the agent loop to a PR-review use case against a real repo\n  Category: Applied\n  Anchor: L-001\n  Rationale: The session used a research example; porting the loop to code review tests generalization.\n\n### Tooling\nE-003 — Package the 3-tool stack as a reusable npm module with tests\n  Category: Tooling\n  Anchor: L-003\n  Rationale: Turns one-off lab code into a durable library the attendee can reuse.\n\n## Traceability Index\n| ID    | Type        | Category \u002F Source  | Evidence                                |\n|-------|-------------|--------------------|-----------------------------------------|\n| L-001 | Learning    | Concept            | claude-agents-workshop.pdf:p4, p5       |\n| L-002 | Learning    | Concept            | claude-agents-workshop.pdf:p11          |\n| L-003 | Learning    | Tool               | claude-agents-workshop.pdf:p17, p18     |\n| L-004 | Learning    | Example            | claude-agents-workshop.pdf:p22-p27      |\n| L-005 | Learning    | Takeaway           | claude-agents-workshop.pdf:p31          |\n| N-001 | Next Step   | Explicit           | claude-agents-workshop.pdf:p40          |\n| N-002 | Next Step   | Referenced         | claude-agents-workshop.pdf:p41          |\n| N-003 | Next Step   | Implicit           | derived from p11                        |\n| E-001 | Enhancement | Depth → L-004      | anchor only                             |\n| E-002 | Enhancement | Applied → L-001    | anchor only                             |\n| E-003 | Enhancement | Tooling → L-003    | anchor only                             |\n```\n\n## Notes on PDF Handling\n\n- The `Read` tool natively supports PDF. For PDFs > 10 pages, the `pages` parameter is **required** (max 20 pages per call). Plan your reads before starting so you cover every page exactly once.\n- If the PDF is image-only (scanned slides with no text layer), the `Read` tool will return little or no text. In that case, stop and report — OCR is out of scope for this skill.\n- Treat the PDF basename (without extension) as the default slug when constructing the output path and when labelling evidence.\n",{"name":5,"tier":25,"description":7,"allowed-tools":49,"argument-hint":50,"disable-model-invocation":45},"Read, Glob, Bash, Write","\u003Cpdf_path> [output_path]","main","b78f8878258c912b66be9d4b56b316515e44c728",".claude\u002Fskills\u002Fapp-workshop-session-pdf-recap-next","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts","https:\u002F\u002Fgithub.com\u002Fcarlosmarte","http:\u002F\u002Fcarlosmarte.me\u002F","2026-08-25 00:53:46",{"source":8,"skillName":5,"commands":59,"prompt":69,"zipAvailable":70},[60,63,66],{"tool":61,"command":62},"npx","npx skills add carlosmarte\u002Fai-agent-artifacts --skill app-workshop-session-pdf-recap-next",{"tool":64,"command":65},"bunx","bunx skills add carlosmarte\u002Fai-agent-artifacts --skill app-workshop-session-pdf-recap-next",{"tool":67,"command":68},"pnpm","pnpm dlx skills add carlosmarte\u002Fai-agent-artifacts --skill app-workshop-session-pdf-recap-next","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fapp-workshop-session-pdf-recap-next\u002FSKILL.md",true,[],[73,100,126,164,190,213],{"skillNo":74,"name":75,"nameZh":6,"description":76,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":77,"categories":78,"keywords":79,"publishedAt":44},"01M6NQ7MN7FKWQE3CKRWKVAC9K","validation_handoff","Forces the agent to conclude its response by generating a strict validation checklist and outlining pending actions, focusing heavily on defensive execution and edge-case testing.","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fvalidation_handoff\u002FSKILL.md",[],[80,81,82,83,84,85,86,87,88,89,90,91,92,93,94,95,96,97,98,99],"handoff","validation","validation handoff","testing","architecture","edge","actions","refactor","execution","system","implementation","checklist","feature implementation","feature","end","case","defensive","hook","at","pending",{"skillNo":101,"name":102,"nameZh":6,"description":103,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":104,"categories":105,"keywords":106,"publishedAt":44},"01M6NQ74C8EKMSABDNN4T6TFBD","project-ployglot-project-scafford","Scaffold a polyglot library project as TypeScript + Python twin packages under `packages\u002Fts\u002F` + `packages\u002Fpy\u002F`, with a shared `SPEC.md`, cross-language `tests\u002Fparity\u002Ffixtures.json`, side-by-side `examples\u002Fsdk\u002F` + `examples\u002Fapi\u002F` documentation, per-package `Makefile`, and a root orchestrator that exposes `make ci`. Open-source friendly — no organization-specific paths or scopes baked in. Use when starting a new polyglot library, porting a single-language library to a twin, or auditing\u002Fregularizing an existing two-language repo against the canonical layout.","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fproject-ployglot-project-scafford\u002FSKILL.md",[],[107,108,109,110,111,112,113,114,115,116,117,118,119,120,121,122,123,34,124,125],"project","project ployglot project scafford","documentation","packages","run","ci","api","open","makefile","Bash","read","orchestrator","only","organization","shared","layout","dry","scope","\u003Cslug>",{"skillNo":127,"name":128,"nameZh":6,"description":129,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":130,"categories":131,"keywords":146,"publishedAt":44},"01M6NQ6VJVPM37YA73T8BCD3VF","playwright-ui-locator","Auto-heal Playwright tests after a frontend developer changes a div's class, id, or other selector. Catches a failing Playwright test, captures the old vs new DOM tree, finds the element that drifted, derives the most stable replacement locator, and proposes the fix as a GitHub pull request (confident heals) or files a GitHub issue (ambiguous \u002F element-removed cases) — never editing the working tree directly. Trigger when Playwright tests fail with \"locator resolved to 0 elements\", \"Timeout waiting for selector\", or after a UI refactor \u002F class rename \u002F Tailwind cleanup broke previously-passing tests.","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fplaywright-test-gh-auto-heal-ui-locator\u002FSKILL.md",[132,135,138,141],{"categoryNo":133,"categoryName":134,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000007","代码审查",{"categoryNo":136,"categoryName":137,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000008","测试与质量",{"categoryNo":139,"categoryName":140,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000009","前端开发",{"categoryNo":142,"categoryName":143,"taxonomyType":144,"taxonomyTypeDesc":145},"01JBSEEDCAT000000000000032","设计师","PROFESSION","适用职业",[147,148,149,150,151,152,116,117,153,154,155,156,157,158,159,160,161,34,162,163],"playwright","ui","playwright ui locator","frontend","cleanup","Tailwind","write","edit","fix","snapshot","auto","never","issue","files","rename","element","request",{"skillNo":165,"name":166,"nameZh":6,"description":167,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":168,"categories":169,"keywords":173,"publishedAt":44},"01M6NQ6BCW05C6Y6VRV0RBQD1V","org-learn-context-window","Distill the current conversation into a reusable command or skill artifact. Use when the user wants to capture a workflow they just executed (create a reusable slash command) or formalize a resolution\u002Fmistake\u002Fcorrection into a reusable skill so the same class of issue never recurs.","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Forg-learn-context-window\u002FSKILL.md",[170],{"categoryNo":171,"categoryName":172,"taxonomyType":21,"taxonomyTypeDesc":22},"01JBSEEDCAT000000000000023","Skill与智能体开发",[42,174,175,176,177,178,179,180,181,182,183,184,36,158,159,185,186,187,188,189],"context","org","window","org learn context window","workflow","create","just","capture","distill","command","user","reusable","conversation","artifact","resolution","skill artifact",{"skillNo":191,"name":192,"nameZh":6,"description":193,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":194,"categories":195,"keywords":197,"publishedAt":44},"01M6NQ66NFPQM1PDBHP1B2G82X","org-agent-skill-creator","Scaffold and validate scope-aware agent skills (company, enterprise, org, team, project, application). Use when the user asks to create a new agent skill, generate a SKILL.md file, scaffold a skill directory, or validate that existing skill directories meet SKILL_SPEC.md (frontmatter, naming rules, structure, scope\u002Ftier).","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Forg-agent-skill-creator\u002FSKILL.md",[196],{"categoryNo":171,"categoryName":172,"taxonomyType":21,"taxonomyTypeDesc":22},[198,199,175,200,201,202,179,203,204,205,206,36,207,208,107,124,209,210,211,212],"skill creator","agent skill","agent skill creator","creator","org agent skill creator","file","Enterprise","team","Rules","structure","new agent","directories","Existing","company","Meet",{"skillNo":214,"name":215,"nameZh":6,"description":216,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":217,"categories":218,"keywords":220,"publishedAt":44},"01M6NQ5NFNK613NR7FM8HPAJ74","app-workshop-recaps-consolidate","Consolidate multiple workshop-session recap Markdown files (typically produced by the workshop-session-pdf-recap-next skill) into a single unified report. Merges and deduplicates learnings across sessions and emits six sections — Consolidated Recap, Follow-ups, Action Items, Next Steps, Possible Enhancements, and Industry \u002F External Viewpoint. The last section supplements the attendees' material with current external perspective via web research. Use when the user has a set of per-session recap files and wants one portfolio-level summary with an outside-in lens (\"consolidate these recaps\", \"merge my workshop notes\", \"cross-session summary\", \"roll up action items\").","https:\u002F\u002Fgithub.com\u002Fcarlosmarte\u002Fai-agent-artifacts\u002Fblob\u002Fmain\u002F.claude\u002Fskills\u002Fapp-workshop-recaps-consolidate\u002FSKILL.md",[219],{"categoryNo":19,"categoryName":20,"taxonomyType":21,"taxonomyTypeDesc":22},[221,25,222,29,223,224,116,225,27,226,227,160,228,26,229,34,230,31,231,232],"consolidate","app workshop recaps consolidate","notes","SET","markdown","portfolio","up","learnings","across","lens","follow-ups","websearch"]