菜单
Create, inspect, extract, render, transform, validate, manifest, and package PDF-based business artifacts.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Build, enhance, fix, or update Tessera playbooks through an interview-first workflow. Use when an agent needs to create a new playbook package from scratch, improve an existing playbook in the workspace, repair a failed or invalid playbook package, update workflow instructions, brainstorm tools/effects/data sources/review gates/final artifacts, generate or modify package files, run validation loops, or prepare portable recipes for Claude Code, Codex, Claude Cowork, Pi Agent, Tessera agents, or any harness that supports skills.
Debug, troubleshoot, and repair Tessera playbook runs from Task UI evidence. Use when a playbook fails, stalls, needs attention, produces a blank or tiny workspace file, writes the wrong artifact, skips a source lookup, cannot read mail/drive/calendar/web data, has suspicious run artifacts/effects/operation logs, or when the user asks to inspect playbook logs, graph run database records, run ids, output files, or execution diagnostics and then fix the playbook package.
Compare options and produce recommendation-ready decision briefs.
Plan multi-step business work with clear checkpoints.
Synthesize research into concise, evidence-grounded takeaways.
Plan, critique, and draft business presentation decks.
| name | pdf-workflows |
| description | Create, inspect, extract, render, transform, validate, manifest, and package PDF-based business artifacts. |
Use this skill when the source or requested output is a PDF: contracts, reports, invoices, board packs, forms, research papers, exports, or scanned documents.
pdf_capabilities before creating, visual review, transforms, or packet work to confirm bundled extraction, creation, transforms, and optional managed engines are available.pdf_inspect to identify page count, text-layer status, scan risk, engine metadata, and warnings.pdf_extract for page-scoped text extraction. Keep page ranges narrow when the user asks about a specific clause, table, figure, or signature block.pdf_render when visual layout, signatures, scans, page appearance, or placement needs review. It uses an optional Tessera-managed PDF render engine when available.pdf_transform only for split, merge, reorder, and rotate operations. It uses Tessera's bundled TypeScript transform engine and writes every transformed PDF to a new output path.pdf_create for simple business PDFs from headings, text, simple tables, page breaks, and embedded workspace images. Include source paths for provenance and validate the created PDF.pdf_validate before relying on a PDF packet and after any created or transformed PDF is produced.pdf_manifest for multi-step packets, created or transformed outputs, handoff, archive, or later business review. Include every material inspect, extract, render, create, transform, and validation result that supports the answer.pdf_inspect: first call for any PDF-specific workflow.pdf_capabilities: readiness check for bundled PDF text extraction, bundled PDF creation/transforms, and optional Tessera-managed render engines.pdf_extract: page-scoped extraction with page markers.pdf_validate: existence, page count, text-layer expectation, pass/fail checks, provenance, and warnings.pdf_render: page-scoped PNG outputs for visual review.pdf_transform: split, merge, reorder, and rotate into new PDF files.pdf_create: simple business PDF creation with headings, text, simple tables, page breaks, embedded workspace images, and source provenance.pdf_manifest: JSON packet manifest for audit, handoff, archive, and future review.workspace_extract: fallback reader for general document extraction when PDF-specific tools are unavailable.Produce a concise extraction, review memo, conversion-ready structure, or file update. Always note scan quality, omitted pages, warnings, validation status, and content that could not be verified from the PDF.