VitaeContext
VitaeGraphGitHub

The career context layer for AI work

Keep your career context.
Reuse it across AI.

VitaeContext is open-source career-context infrastructure for AI agents and career systems. It helps a person maintain reusable Career Context in Markdown, then provides focused agent skills and a Model Context Protocol (MCP) server for grounded CVs, profiles, applications, and professional work across any workspace.

The file stays under your control until you supply it to an external provider. VitaeContext structures supplied facts and evidence boundaries; it does not independently verify every claim or guarantee outcomes.

Install VitaeContextSee context example

  • User-owned Markdown
  • Reusable context
  • Evidence boundaries
CV
GitHub
notes
your material
MD career-context.md

# Career context

## Facts
## Goals
## Experience
## Projects
## Claims to avoid
reusable source of truth
LinkedIngrounded
CVgrounded
Portfoliogrounded
professional tasks
Maintain the context once. Reuse it instead of rebuilding your history in every chat.

Context before automation

Career systems need good career information.

Systems that create applications, tailor CVs, or handle other career tasks need professional career information. If that information is incomplete or wrong, the whole workflow starts from a weak foundation.

01Career materialCVs, projects, experience, goals, proof
02 · VitaeContextCareer ContextUser-owned, maintained Markdown
03Any career systemAgents, CV tools, application workflows, drafts

VitaeContext works before the task. It is for people maintaining career information and for developers or agents that need precise context before professional work.

Why context first

Career information gets scattered.

CVs, profiles, project notes, and old chats often disagree or miss useful detail. A maintained context file gives each new task the same starting facts and direction.

CV-final-v7.pdf“Led product work…”last updated?
LinkedIn About“Building useful AI tools…”different wording
project-notes.mdLaunch date + proof linknot in profile
new AI chat“Tell me about yourself.”start again
MD
LOCAL · EDITABLE · PORTABLEcareer-context.md
Supported facts
identity, roles, projects, metrics
Stated direction
goals, target roles, interests
Proof
links, evidence, source notes
Boundaries
constraints and claims to avoid

Context supplies facts. Skills supply method.

Use the same context for different tasks.

Load one focused skill for the work in front of you. It adds task-specific structure and guidance without becoming another career record.

Reusable input
MDcareer-context.mdfacts + goals + evidence
Focused method
$vitaecontext-[skill]platform rules + workflow
Task output
✓Grounded workadapted, not invented

Context can also ground professional messages, interview preparation, and other career-related drafts in any AI chat. Choose a workflow by task if you are not sure where to start.

Model Context ProtocolStateless Server

Query private career facts across any repository.

VitaeContext includes a stateless MCP server that connects your private Career Context and VitaeGraph directly to AI coding tools. Agents across all project workspaces can query verified career facts on demand without copying files into git repositories.

Zero file pollutionCareer files stay in ~/.vitaecontext/. Never committed to project repositories.
Stdio JSON-RPC 2.0Zero external dependencies. Instant startup via npx -y vitaecontext-mcp.
Read MCP Guide 11 Supported Providers
stdio · json-rpc 2.0 · stateless
$ npx -y vitaecontext-mcp
Server ready · listening on stdio
TOOLget_career_context{ "for": "github" }
TOOLsearch_vitaegraph{ "type": "project" }
RESOURCEcareer-context://current~/.vitaecontext/
PROMPTcv_tailoringtargetRole

Keep claims supported

Change the writing, not the facts.

Rewriting often becomes more confident with every pass. VitaeContext tells agents to preserve supported claims, keep goals separate from verified experience, and flag evidence gaps before strengthening copy.

  • 01
    Use supported claimsBase work on supplied context and sources.
  • 02
    Keep evidence boundariesDo not turn intent or inference into experience.
  • 03
    Adapt for the surfaceChange structure and language, not facts.
SUPPORTED SOURCE

Built and maintains an open-source toolkit of portable agent skills.

Repeated rewrite

“Created a revolutionary career growth platform.”

Evidence disappeared
Context + skill

“Maintains an open-source toolkit of portable agent skills.”

✓ Claim stays supported
VitaeGraphpowered by VitaeContext

Use separate records when one file is not enough.

VitaeGraph is the deeper structured-context module: local, human-readable Markdown records for projects, roles, education, courses, thesis work, certifications, awards, and publications, plus relationships between them.

Context fileFast facts and repeated positioning
VitaeGraphDetailed records and relationships
Explore VitaeGraph View graph commands
~/.vitaecontext/vitaegraph/
Graph summaryVITAEGRAPH.md

Identity, direction, and links to detailed records.

Markdown canonical

Detailed records

ED
education/degree.mdcontains thesis.md and course.md
PR
projects/search-tool.mdrelated to role and publication records
EX
experience/role.mdrelated to project and publication
contains relatedSmall records. Explicit links.

Get started

Install, create context, choose a task.

VitaeContext installs as readable skills. Career context remains in files you choose and control until you supply it to an AI provider.

  1. 01

    Install or connect MCP

    npx vitaecontext install --provider codex
# or stdio MCP: npx -y vitaecontext-mcp

    Supports 11 agent targets including Cursor, Windsurf, Claude Code, Codex, Roo Code, and Grok.

    All providers
  2. 02

    Initialize private context

    npx vitaecontext context init --output ~/.vitaecontext/name-surname-career-context.md

    Initialize your private Markdown file, then let the context builder organize trusted material.

    View example
  3. 03

    Run grounded tasks

    $vitaecontext-github
Audit my profile using my career context file.

    Reference the context, invoke the focused skill or MCP tool, and ask for grounded output.

    Usage guide
OPEN SOURCE · MIT LICENSEGive each task a clear starting point.
Install VitaeContextView source on GitHub