nav-marker_skill

This skill creates context checkpoints to preserve conversation state before breaks or risky changes, enabling seamless restoration later.
  • Python

142

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill alekspetrov/navigator --skill nav-marker

  • SKILL.md10.1 KB

Overview

This skill creates context save points (markers) that preserve conversation state so you can pause work and resume later without re-explaining. It captures conversation summary, technical decisions, current focus, and a Theory-of-Mind section to restore mental context and reduce onboarding friction when resuming.

How this skill works

When invoked, the skill ensures a local marker directory exists, determines or prompts for a marker name and optional note, then synthesizes a structured markdown marker capturing the last part of the conversation. It actively extracts user intent, corrections, belief state, files touched, loaded documentation, next steps, and verification metadata, then writes and verifies the marker file and logs creation details.

When to use it

  • When a user says: "save my progress", "create checkpoint", or "mark this point"
  • Before taking a break (lunch, end of day) or stepping away
  • Before a risky refactor or experimental change
  • Before compacting or clearing conversation context
  • When reaching a milestone and you want a clear restore point

Best practices

  • Use descriptive names (e.g., before-api-refactor, lunch-break) to make restores intuitive
  • Include a short note to capture context that may not be obvious later
  • Avoid creating markers for very fresh contexts (< 5 messages)
  • Capture ToM fields (Primary goal, preferences, corrections) — do not leave them blank
  • Verify file creation and checksum to ensure marker integrity

Example use cases

  • User: "Save my progress, taking lunch" → creates lunch-break marker to resume later
  • User: "Mark this before I refactor routing" → snapshot state to roll back if needed
  • End-of-day checkpoint: capture what’s done, blockers, and next steps for tomorrow
  • Automatic marker before context compaction so knowledge is preserved across clears
  • Create a marker before trying an experimental approach to protect a known-good state

FAQ

A marker captures a conversation summary, loaded documentation, files modified, technical decisions, next steps, ToM (user intent, preferences, corrections), and belief state.

How is the marker verified?

After writing the file the skill checks the file exists and is non-empty, computes a checksum, logs creation to a central marker log, and reports verification details.

When should I not create a marker?

Do not create markers for very fresh sessions (< 5 messages) or when the user only wants to list or inspect existing markers.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational