Dual-context skill: Personal TELOS reads and updates goals, beliefs, narratives, strategies, and more with timestamped backups; Project TELOS analyzes .md/.csv
复制下面这句话,粘贴给 Claude Code、Codex、Cursor 等 AI 编程工具,它会读取安装说明并在你确认后完成安装。
请阅读 https://ai.atlankj.com/install/asset/gh-telos-62c3c05ba736 ,按照其中的说明把「Telos」安装到你(当前 AI 工具)中。执行前先告诉我将运行的命令和写入的位置,等我确认。
查看 AI 将读取的安装说明正在读取 GitHub 原文…
内容来自 GitHub 原始文件,由原作者维护。在 GitHub 查看
You MUST send this notification BEFORE doing anything else when this skill is invoked.
Send voice notification:
curl -s -X POST http://localhost:31337/notify \
-H "Content-Type: application/json" \
-d '{"message": "Running the WORKFLOWNAME workflow in the Telos skill to ACTION"}' \
> /dev/null 2>&1 &
Output text notification:
Running the **WorkflowName** workflow in the **Telos** skill to ACTION...
This is not optional. Execute this curl command immediately upon skill invocation.
Reads and updates two kinds of context. Personal TELOS: the principal's life context — goals, beliefs, wisdom, books, movies, challenges, narratives, strategies, mission, models, predictions, traumas, frames, lessons, wrong-beliefs — at ~/.claude/LIFEOS/USER/TELOS/, updated through the Update workflow with timestamped backups. Project TELOS: analyzes a directory of .md/.csv files to extract dependency chains (PROBLEMS→GOALS→STRATEGIES→PROJECTS), bottlenecks, alignment, and progress, then produces a McKinsey-style report, n=24 narrative points, or a Next.js dashboard built by parallel engineers.
Life context and project state both sprawl across many files, and acting on them by hand is error-prone. Personal goals, beliefs, and lessons drift out of date, and hand-editing them risks losing history or corrupting structure. Project documentation hides the dependency chains, bottlenecks, and misalignments that decide whether the work is on track — you can read every file and still not see how PROBLEMS connect to GOALS connect to PROJECTS. This skill gives both a single safe way in: structured updates with backups for the personal side, and automated relationship analysis plus rendered outputs for the project side.
TELOS (Telic Evolution and Life Operating System) is a context-gathering system with two applications:
~/.claude/LIFEOS/USER/TELOS/The skill detects which context a request means (see Context Detection below), then routes to the right workflow. Personal updates always go through the Update workflow so backups and changelog entries happen automatically. Project analysis scans the target directory, builds a relationship graph, and renders the output format you asked for.
When executing a workflow, output this notification directly:
Running the **WorkflowName** workflow in the **Telos** skill to ACTION...
| Workflow | Trigger | File |
|---|---|---|
| Update | "add to TELOS", "update my goals", "add book to TELOS" | Workflows/Update.md |
| InterviewExtraction | "extract content", "extract interviews", "analyze interviews" | Workflows/InterviewExtraction.md |
| CreateNarrativePoints | "create narrative", "narrative points", "TELOS report", "n=24" | Workflows/CreateNarrativePoints.md |
| WriteReport | "write report", "McKinsey report", "create TELOS report", "professional report" | Workflows/WriteReport.md |
Note: For general project analysis, dashboards, dependency mapping, and executive summaries, the skill handles these directly without a separate workflow file.
Example 1: Update personal TELOS
User: "add Project Hail Mary to my TELOS books"
--> Invokes Update workflow
--> Creates timestamped backup of BOOKS.md
--> Adds book entry with formatted metadata
--> Logs change in updates.md with timestamp
Example 2: Analyze project with TELOS
User: "analyze ~/Projects/MyApp with TELOS"
--> Scans all .md and .csv files in directory
--> Extracts entities, relationships, dependencies
--> Returns analysis with dependency chains and progress metrics
Example 3: Build project dashboard
User: "build a dashboard for TELOSAPP"
--> Launches up to 16 parallel engineers
--> Creates Next.js dashboard with shadcn/ui + Aceternity
--> Returns interactive dashboard with dependency graphs, metrics cards, progress tables
Example 4: Generate narrative points
User: "create TELOS narrative for Acme Corp, n=24"
--> Invokes CreateNarrativePoints workflow
--> Analyzes TELOS context (situation, problems, recommendations)
--> Returns 24 crisp bullet points (8-12 words each)
--> Output is slide-ready for presentations or customer briefings
Example 5: Generate McKinsey-style report
User: "write a TELOS report for Acme Corp"
--> Invokes WriteReport workflow
--> First runs CreateNarrativePoints to generate story content
--> Maps narrative to McKinsey report structure
--> Generates web-based report with professional styling
--> Output at {project_dir}/report - run `bun dev` to view
--> White background, subtle Tokyo Night Storm accents
--> Includes: cover page, executive summary, findings, recommendations, roadmap
How {DA_IDENTITY.NAME} determines which TELOS context:
| User Request | Context | Location |
|---|---|---|
| "my TELOS", "my goals", "my beliefs", "add to TELOS" | Personal TELOS | ~/.claude/LIFEOS/USER/TELOS/ |
| "Alma", "TELOSAPP", "analyze [project]", "dashboard for" | Project TELOS | User-specified directory |
| "analyze ~/path/to/project" | Project TELOS | Specified path |
CRITICAL PATH: All personal TELOS files are located at:
~/.claude/LIFEOS/USER/TELOS/
Personal TELOS lives in the CORE USER directory, NOT directly under the Telos skill directory.
All files located in ~/.claude/LIFEOS/USER/TELOS/:
# View specific file
read ~/.claude/LIFEOS/USER/TELOS/GOALS.md
read ~/.claude/LIFEOS/USER/TELOS/BELIEFS.md
# View recent updates
read ~/.claude/LIFEOS/USER/TELOS/updates.md
CRITICAL: Never manually edit. Use the Update workflow.
Workflow: Workflows/Update.md
The workflow provides:
Valid files for updates: BELIEFS.md, BOOKS.md, CHALLENGES.md, FRAMES.md, GOALS.md, LEARNED.md, MISSION.md, MODELS.md, MOVIES.md, NARRATIVES.md, PREDICTIONS.md, PROBLEMS.md, PROJECTS.md, STRATEGIES.md, TELOS.md, TRAUMAS.md, WISDOM.md, WRONG.md
For any project directory, TELOS provides:
Flexible file discovery - no required structure:
# User specifies directory
"Analyze ~/Projects/<your-org-tree>"
--> {DA_IDENTITY.NAME} scans for .md and .csv files anywhere in tree
# {DA_IDENTITY.NAME} automatically finds all .md and .csv files regardless of structure
Point the skill at a project directory (named project, explicit path, or common location). A good analysis surfaces the dependency chains PROBLEMS→GOALS→STRATEGIES→PROJECTS, the bottlenecks blocking progress, whether projects align with stated objectives, progress/completion metrics, and risk areas (overdue or blocked work) — grounded in what the files actually say, never cached assumptions.
Discover the source files:
find $TARGET_DIR -type f \( -name "*.md" -o -name "*.csv" \)
Deliver in whatever format was asked: markdown report (Mermaid diagrams), interactive web dashboard, JSON export, or executive summary.
The deliverable is an interactive Next.js dashboard that renders the analysis against the stack and design tokens below. Start from the shipped scaffold at DashboardTemplate/ (working Next.js app: file browser, markdown/CSV rendering, dependency views — see its README for the copy-and-run steps) rather than from scratch, then adapt pages to the analysis. Fan the build out across parallel agents however the work splits cleanly — data-parsing lib, shared components, per-page views, theme, integration — running independent pieces at once. (Wired per public issue #1555, @tzioup — the template previously shipped unreferenced.)
Tech Stack:
Features:
Design tokens:
--background: #ffffff
--foreground: #1a1b26
--primary: #2e7de9
--accent: #9854f1
--destructive: #f52a65
--success: #33b579
--warning: #f0a020
Standard Project TELOS Structure (auto-detected):
Note: Files are optional. TELOS adapts to whatever exists.
Available Visualizations:
Personal TELOS:
Project TELOS:
After completing any workflow, append a single JSONL entry:
echo '{"ts":"'$(date -u +%Y-%m-%dT%H:%M:%SZ)'","skill":"Telos","workflow":"WORKFLOW_USED","input":"8_WORD_SUMMARY","status":"ok|error","duration_s":SECONDS}' >> ~/.claude/LIFEOS/MEMORY/SKILLS/execution.jsonl
Replace WORKFLOW_USED with the workflow executed, 8_WORD_SUMMARY with a brief input description, and SECONDS with approximate wall-clock time. Log status: "error" if the workflow failed.