Who it is forA project lead or developer picking up a long AI-assisted project whose goal and decisions have drifted into old chat history.
The problemGoals drift, old requirements reappear as facts, and decisions are lost inside chat.
The outcomeA status view with the goal, current stage, evidence freshness, blocker, and one next action, plus a checkpoint another task can continue from.
What you need first
- Git
- Codex with local plugin support
- Python 3 for the requirement-ledger CLI, installed with uv or pip
- A local project folder to select as the target
Steps
- Run
codex plugin marketplace add "$PWD"from the GPT 6 Astra Skill checkout. - Run
codex plugin add requirement-ledger@requirement-ledger-local. - Select the project, start a fresh Codex task, and ask: "Hi Jarvis, take over this project. Recover its goal, current stage, blocker, and one next action. Do not modify it."
- If the Skill is not selected automatically, retry once with
$requirement-ledger-workflow Hi Jarvis, take over this selected project.
Prompt for your agent
Using the Jarvis workflow from the gpt-6-astra-skill repo, take over the project at <project-path> without modifying it. Return its goal, current stage, blocker, and one next action with the required authority. Verify that no project files changed before reporting the result.
Caveat. The first screen is a host-selected, unbound Codex decision aid, not a CLI-created source pack or final report.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for adand-91/gpt-6-astra-skill and "Recover a stalled project's goal, stage, and blocker"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer