Agent Skillskbanc85/claudia › file-document

file-document

GitHub

用于存储文档、邮件或文本以便后续参考,支持实体链接和来源追踪。适用于用户保存需保留或引用的内容,通过元数据收集和路由实现结构化归档与检索。

plugins/claudia/skills/file-document/SKILL.md kbanc85/claudia

Trigger Scenarios

用户要求保存邮件或文档 用户粘贴内容希望后续引用 用户上传文件如PDF或合同 用户分享值得保留的研究或网页内容

Install

npx skills add kbanc85/claudia --skill file-document -g -y
More Options

Non-standard path

npx skills add https://github.com/kbanc85/claudia/tree/main/plugins/claudia/skills/file-document -g -y

Use without installing

npx skills use kbanc85/claudia@file-document

指定 Agent (Claude Code)

npx skills add kbanc85/claudia --skill file-document -a claude-code -g -y

安装 repo 全部 skill

npx skills add kbanc85/claudia --all -g -y

预览 repo 内 skill

npx skills add kbanc85/claudia --list

SKILL.md

Frontmatter
{
    "name": "file-document",
    "description": "Store a document, email, or text for future reference with entity linking and provenance. See also: `ingest-sources` for multiple files; `capture-meeting` if it's meeting notes; `summarize-doc` to get a summary without filing.",
    "effort-level": "medium"
}

File Document

Store any document, email, or text for future reference with proper entity linking and provenance tracking.

Usage

/file-document or natural language:

  • "Save this email"
  • "File this for later"
  • "Keep this document"
  • "Store this transcript"

When to Use

This command is for ad-hoc document capture. Use it when:

  • User shares an email they want to preserve
  • User pastes content they want to reference later
  • User uploads a document (PDF, contract, proposal)
  • User shares research or web content worth keeping
  • Anything the user might want to cite or review later

Note: For meeting transcripts specifically, use /capture-meeting which includes extraction.

Quick Flow

1. Identify the Content

Ask if not obvious:

  • "What is this? (email, document, transcript, research, other)"
  • "Who is this about or from?"

2. Gather Metadata

Determine:
├── source_type: gmail, transcript, upload, capture
├── filename: YYYY-MM-DD-[entity]-[topic].md
├── about: [list of entity names mentioned]
└── summary: One-line description

3. File It

claudia memory document store \
  --filename "Descriptive name" \
  --source-type "gmail|transcript|upload|capture" \
  --summary "Brief description" \
  --about "entity1,entity2" \
  --project-dir "$PWD" \
  < content.md

(Pipe the FULL raw text via stdin; never summarize)

4. Confirm

**Filed**

**File:** [filename]
**Location:** [storage path]
**Linked to:** [entity names]
**Summary:** [one-line summary]

The document is now searchable and linked to [entities].
You can find it later with:
- "Show me documents about [entity]"
- "Find the email from [person]"
- claudia memory document search --entity "[name]" --project-dir "$PWD"

Source Types

Type Use When Example
gmail Email content "Here's an email from Sarah..."
transcript Meeting notes/recordings "Notes from today's call..."
upload PDFs, contracts, formal docs "Here's the contract..."
capture Research, web content, misc "I found this article..."

File Routing

Documents are automatically routed to entity-aware folders:

If about a person:
  people/sarah-chen/emails/2026-02-04-proposal.md
  people/sarah-chen/documents/2026-02-04-contract.pdf

If about an organization:
  clients/acme-corp/emails/2026-02-04-partnership.md

If about a project:
  projects/rebrand/documents/2026-02-04-brief.md

If no entity linked:
  general/documents/2026-02-04-misc.md

Examples

Email

User: "Here's an email from Jim about the partnership terms. Save it."

claudia memory document store \
  --filename "2026-02-04-jim-ferry-partnership.md" \
  --source-type "gmail" \
  --summary "Jim Ferry re: partnership terms and next steps" \
  --about "Jim Ferry" \
  --project-dir "$PWD" < email.md

Response:
"Filed Jim's email about partnership terms.
Saved to: people/jim-ferry/emails/2026-02-04-partnership.md
You can find it later by asking about Jim's documents."

Research

User: "Here's some info I found about competitor pricing. Keep this."

claudia memory document store \
  --filename "2026-02-04-competitor-pricing-research.md" \
  --source-type "capture" \
  --summary "Competitor pricing analysis notes" \
  --project-dir "$PWD" < research.md

Response:
"Filed your competitor pricing research.
Saved to: general/documents/2026-02-04-competitor-pricing-research.md"

Contract

User: "Save this contract from Acme Corp"

claudia memory document store \
  --filename "2026-02-04-acme-corp-contract.md" \
  --source-type "upload" \
  --summary "Service agreement with Acme Corp" \
  --about "Acme Corp" \
  --project-dir "$PWD" < contract.md

Response:
"Filed the Acme Corp contract.
Saved to: clients/acme-corp/documents/2026-02-04-contract.md"

Linking to Memories

If you also extract facts from the document:

  1. File the document first (get document_id from JSON response)
  2. Extract memories using claudia memory save or claudia memory batch
  3. Call claudia memory document store --memory-ids "id1,id2,..." to link provenance

This creates the chain: memory -> document -> file on disk.

Quality Checklist

  • Full content preserved (not summarized)
  • Descriptive filename with date
  • Correct source_type selected
  • Entities identified and linked
  • Summary is accurate and searchable
  • User knows where to find it later

Tone

  • Quick and efficient
  • Confirm what was saved
  • Tell them how to find it later
  • Don't over-explain the system

Version History

  • 80253c0 Current 2026-08-20 06:21

Same Skill Collection

codex/skills/build-team/SKILL.md
codex/skills/claudia-core/SKILL.md
codex/skills/databases/SKILL.md
codex/skills/onboarding/SKILL.md
plugins/claudia/skills/auto-research/SKILL.md
plugins/claudia/skills/brain-monitor/SKILL.md
plugins/claudia/skills/brain/SKILL.md
plugins/claudia/skills/build-team/SKILL.md
plugins/claudia/skills/capture-meeting/SKILL.md
plugins/claudia/skills/claudia-core/SKILL.md
plugins/claudia/skills/client-health/SKILL.md
plugins/claudia/skills/databases/SKILL.md
plugins/claudia/skills/deep-context/SKILL.md
plugins/claudia/skills/diagnose/SKILL.md
plugins/claudia/skills/draft-reply/SKILL.md
plugins/claudia/skills/feedback/SKILL.md
plugins/claudia/skills/financial-snapshot/SKILL.md
plugins/claudia/skills/fix-duplicates/SKILL.md
plugins/claudia/skills/follow-up-draft/SKILL.md
plugins/claudia/skills/growth-check/SKILL.md
plugins/claudia/skills/inbox-check/SKILL.md
plugins/claudia/skills/ingest-sources/SKILL.md
plugins/claudia/skills/map-connections/SKILL.md
plugins/claudia/skills/meditate/SKILL.md
plugins/claudia/skills/meeting-prep/SKILL.md
plugins/claudia/skills/memory-audit/SKILL.md
plugins/claudia/skills/memory-health/SKILL.md
plugins/claudia/skills/morning-brief/SKILL.md
plugins/claudia/skills/new-person/SKILL.md
plugins/claudia/skills/new-workspace/SKILL.md
plugins/claudia/skills/onboarding/SKILL.md
plugins/claudia/skills/pipeline-review/SKILL.md
plugins/claudia/skills/research/SKILL.md
plugins/claudia/skills/skill-router/SKILL.md
plugins/claudia/skills/summarize-doc/SKILL.md
plugins/claudia/skills/weekly-review/SKILL.md
plugins/claudia/skills/what-am-i-missing/SKILL.md
template-v2/.claude/skills/auto-research/SKILL.md
template-v2/.claude/skills/client-health/SKILL.md
template-v2/.claude/skills/databases/SKILL.md
template-v2/.claude/skills/deep-context/SKILL.md
template-v2/.claude/skills/feedback/SKILL.md
template-v2/.claude/skills/financial-snapshot/SKILL.md
template-v2/.claude/skills/fix-duplicates/SKILL.md
template-v2/.claude/skills/new-workspace/SKILL.md
template-v2/.claude/skills/pipeline-review/SKILL.md
template-v2/.claude/skills/research/SKILL.md
plugins/claudia/skills/wiki/SKILL.md
template-v2/.claude/skills/wiki/SKILL.md

Metadata

Files
0
Version
80253c0
Hash
fe459258
Indexed
2026-08-20 06:21

inicio - Wiki
Copyright © 2011-2026 iteam. Current version is 2.155.2. UTC+08:00, 2026-08-22 00:17
浙ICP备14020137号-1 $mapa de visitantes$