- Home
- Skills
- Technical Documentation
- Context Waypoint
Context Waypoint
Portable, Markdown-based project memory to maintain continuity and technical context across AI coding sessions.
Free
Works with the AI Agents you already use
Context Waypoint
Example session with this skill installed
We just finished migrating the auth flow to Auth0. Initialize the waypoint and record this decision including the fact we rejected Firebase for OIDC compatibility reasons.
- Read your context and instructions
- Compiled the context waypoint
I have initialized the .context-waypoint/ directory and created the foundational project records. I documented the Auth0 migration and the specific rationale for rejecting Firebase to prevent redundant evaluations in future sessions.
Project Context Initialized
| File | Status | Summary |
|---|---|---|
CURRENT.md | Created | Set objective to Auth0 migration cleanup and OIDC hardening. |
CONSTRAINTS.md | Created | Added OIDC compliance requirement for all auth modules. |
decisions/0001-auth0-migration.md | Recorded | Migration from legacy auth to Auth0; Firebase rejected (OIDC). |
Decision Record: 0001-auth0-migration.md
# Decision: Migrate to Auth0 for Identity Management
* Status: Accepted
* Date: 2023-10-27
* Deciders: AI Agent, User
## Context and Problem Statement
The legacy custom auth flow was insecure and lacked OIDC support. We needed a managed solution that integrates with our existing React frontend and Node.js backend.
## Decision Outcome
Chosen Option: Auth0
* Rationale: Full OIDC support, superior documentation for our stack, and built-in MFA.
* Consequences: Requires `auth0-js` and `express-oauth2-jwt-bearer` dependencies.
## Pros and Cons of Alternatives
### Firebase Auth
* Good: Fast initial setup.
* Bad: Rigid OIDC provider support made it incompatible with our enterprise requirements.
Next steps
- Review
CONSTRAINTS.mdto ensure the OIDC guardrails match your security policy. - Commit the
.context-waypoint/directory to your repository to share this state with other agents. - Use the
handoffcommand when you are ready to switch tasks.
Connects securely to your tools. The creator never sees your data.
About this skill
AI coding assistants lose context between sessions, leading to repeated mistakes and wasted tokens on redundant explanations. This skill solves project amnesia by maintaining a portable, Markdown-based memory layer directly within your repository.
What it does
- State Persistence creates durable records of current objectives, blockers, and next actions in
CURRENT.md. - Constraint Tracking maintains a
CONSTRAINTS.mdto ensure guardrails and architectural rules are never forgotten. - Decision Logging captures the rationale behind technical choices in ADR-style records to prevent circular debates.
- Session Summarization distills meaningful work and verified conclusions without bloating the context window with chat history.
- Handoff Management generates structured handoff files that allow different agents or models to resume unfinished tasks instantly.
How it works
- Bootstrap the repository by running an initialization command to create the
.context-waypoint/structure and templates. - Synchronize state by updating the
CURRENT.mdfile whenever the working objective or project status shifts. - Capture major architectural or technical changes by generating a new record in the
decisions/directory. - Handoff work before pausing a session, creating a specific bridge for the next agent to ingest.
Frameworks & tools
This skill is platform-agnostic and uses standard Markdown. It is designed for Git-based workflows, Claude Code, Cursor, and other agentic IDEs that can read and write to the local filesystem.
Why this beats prompting it yourself
Manual prompting for project memory is inconsistent and consumes massive token overhead. This skill provides a rigorous schema and operating model that forces the AI to store only high-signal evidence, ensuring your context window stays lean while your project memory remains accurate.
Use cases
- Resuming a complex refactor after a weekend break or context switch.
- Onboarding a new AI agent to a codebase with specific architectural constraints.
- Documenting why a specific library was rejected to prevent the AI from suggesting it again.
- Bridging context between a specialized coding agent and a general-purpose architect agent.
Known limitations
Does not automate Git commits or pushes. Requires the agent to have filesystem write permissions. Does not include a background daemon for automatic real-time logging.
Requires filesystem write access. No automatic Git integration. The user must explicitly trigger state updates. Not intended for secrets or raw log storage.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
1 use
Times used to date
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 1 month ago
- Free with an Agensi account