
Auto-collect Feishu/DingTalk data, generate Work+Persona dual-dimension Skill
Colleague-skill is an interactive team socialization and knowledge-cloning skill on EasyClaw. It integrates with enterprise collaboration data — securely parsing public chat history logs, technical specifications, and project memos from platforms like Feishu (Lark) or DingTalk — to construct a reusable digital profile of a team member. This profile captures both their professional expertise (their specific project knowledge and technical workflows) and their personal style (their tone of voice, analytical approaches, and communication preferences).
The skill is built for fast-growing startups, remote-first agencies, and corporate teams who want to build a persistent, easily accessible directory of team expertise — preventing knowledge silos, smoothing employee transitions, and automating routine peer consultations.
The expected outcome is a structured, dual-dimension Digital Colleague Profile: a persistent workspace module containing that teammate's compiled domain knowledge alongside a conversational toggle that lets you interact with their cloned analytical style.
1. Source data ingestion. Provide public text-based materials related to a colleague — such as past chat logs, project reports, design specifications, or PDF manuals they authored.
2. Dual-dimension extraction. The skill parses the ingested materials across two distinct layers:
- Work Dimension (Knowledge): Compiles the colleague's hard domain expertise, system maps, codebase preferences, and specific project histories.
- Persona Dimension (Style): Analyzes and abstracts their communication patterns, favorite frameworks, vocabulary habits, and general analytical approach.
3. Profile synthesis. It structures these extracted data points into a clean, unified Digital Profile directory, organizing knowledge blocks by category (e.g., "Legacy Database Architecture" or "Marketing Copy Guidelines").
4. Interactive companion integration. The skill maps these dimensions into a conversational toggle, enabling other team members to query the profile for specific factual answers or discuss project directions using the colleague's cloned analytical style.
5. Continuous material appending. You can continuously upload new PDFs, transcripts, or meeting logs to update specific colleague profiles, ensuring the knowledge directory remains current as projects evolve.
- Dual-dimension cloning: Separates hard project knowledge (Work) from communication style (Persona) for accurate profiling.
- Enterprise chat parsing: Extracts useful patterns and project histories from raw Feishu/DingTalk chat logs.
- PDF knowledge extractor: Automatically structures technical specifications and project files into searchable folders.
- Tone-of-voice synthesizer: Analyzes and replicates specific communication habits and vocabulary preferences.
- Dynamic profile appending: Easily update existing colleague profiles by uploading new memos or logs.
- Frictionless query interface: Let team members ask questions and receive answers formatted in that colleague's style.
1. Preserving legacy system knowledge during employee transitions
A senior backend developer ("Zhang San") who designed your company's legacy database is leaving the company. You compile their public technical design specs, database diagrams, and Slack Q&A threads, and feed them into Colleague-skill. It builds the "Zhang San Cloned Skill." The incoming developer can ask the cloned skill complex technical questions — "where are the database user-tables indexed?" — and receive Zhang San's exact architectural reasoning, preventing key knowledge loss.
2. Scaling design reviews across a remote team
An agency's lead creative director has a highly specific design style and feedback pattern, but is overwhelmed with review meetings. The agency team clones their review style using past feedback logs. Designers can submit their draft mockups to the cloned skill to receive a "pre-review" check, identifying layout or font-pairing issues in the creative director's exact style before booking a face-to-face meeting.
3. Onboarding new team members efficiently
A newly hired marketer needs to understand the brand's positioning and past campaign results. Instead of booking hours of onboarding calls, they query the cloned "Marketing Lead Profile" in EasyClaw — reading through campaign summaries, past copy guidelines, and target audience definitions, accelerating their onboarding timeline.
4. Resolving cross-functional communication gaps
A product manager wants to discuss a feature draft with the engineering lead, but struggles to anticipate their technical objections. They query the cloned "Engineering Lead Profile" with their draft, receiving feedback structured around the engineering lead's favorite technical considerations (e.g., server load, API stability), allowing them to refine their pitch before the meeting.
5. Automated client-outreach copy review
A sales team wants to ensure their outbound emails match the warm, consultative tone of their top-performing sales representative. They clone the representative's copy style. Sales reps can run their email drafts through the cloned skill to get vocabulary adjustments and structural polishes that match the high-converting writing style.
An agency founder wants to clone their lead copywriter's ("Marie") unique copy-review style to help junior writers iterate faster.
1. They open EasyClaw and activate Colleague-skill.
2. They run: *"Clone colleague: distill Marie's copywriting and review style into a skill. Here are her past copy guides and feedback logs."*
3. The skill parses the texts, extracts the core rules (e.g., "always lead with the customer's problem," "avoid corporate buzzwords"), and models her warm but direct feedback tone.
4. It outputs the Marie Cloned Profile containing the Work and Persona parameters.
5. A junior writer uploads their draft landing page copy.
6. The cloned skill reviews the draft, highlights the buzzwords Marie would hate, and provides rewrites in her exact copy-editing style.
Marie's unique copy-editing expertise scaled across the entire writing team.
Clone colleague: distill Zhang San's work style into a skill -> Generate reusable Work+Persona dual-dimension Skill
Append material: add new chat logs to this skill -> Missing colleague name, please specify: which colleague to update?
Upload file: add technical specs from this PDF -> Auto-extract and update to corresponding colleague Skill
Yes, when used ethically for professional knowledge sharing. The skill is designed to compile publicly accessible, business-owned project files, chat logs, and manuals to preserve institutional knowledge. It does not access private personal data. We recommend obtaining clear team consent and alignment before initializing colleague profiles.
The skill can ingest and parse text-based exports from any major enterprise collaboration platform, including Feishu (Lark), DingTalk, Slack, Microsoft Teams, and standard email or PDF manuals.
The style synthesizer analyzes linguistic patterns: average sentence length, preferred vocabulary, standard layout structure, and typical tone (e.g., warm, direct, analytical, or formal), mapping these into its conversational response guidelines.
Yes. Once a colleague profile is generated and verified, you can integrate it with your enterprise Feishu bot using ClaudeIM to allow team members to query the clone directly inside their active chat channels.
To update an existing profile, specify: "Append technical specs from this PDF to [Colleague Name]'s profile." The skill parses the new file, extracts the new knowledge blocks, and merges them into the existing directory.
Yes. All profile files, ingested texts, and access models are stored locally in your workspace directories. You can securely delete a colleague's profile JSON and associated archives in one click.
The skill generates a clean, structured JSON file containing the colleague's knowledge base directories, style guidelines, and interaction rules, saved securely under `public/data/skills/clones/`.
During data ingestion, the skill identifies recurring company-specific terms and acronyms, parses their context from the files, and compiles an internal jargon index to ensure accurate communication.
If the cloned colleague is a senior developer and you ingest their technical specifications and coding guidelines, the clone can output code blocks aligned with their specific architectural preferences and styling guidelines.
No. All data extraction, synthesis, and directory writes run headlessly in the background, keeping your active conversation workspace fast and responsive.
Browse more in General Tools or all skills.
Get EasyClaw, add this skill, and start building AI agent workflows in minutes.
Get EasyClaw Free →