Write a cold-start task briefing file to /tmp/<descriptive-slug>-task.md.
npx skills add https://github.com/jcdendrite/claude-config --skill brief-jcdendrite
Please help me install this Agent Skill. Skill: brief Source: https://github.com/jcdendrite/claude-config/tree/main/claude/.claude/skills/brief Command: npx skills add https://github.com/jcdendrite/claude-config --skill brief-jcdendrite
This Skill turns an abandoned or settled piece of work into a cold-start briefing that a fresh session can safely execute without relying on prior conversation context.
Ask the agent to run the brief skill to create a new execution-ready task file for a settled, abandoned PR or ticket.
High-intent search queries and answers about installing and using this skill.
To create a task briefing for a fresh agent session, use the brief skill to generate a cold-start task file at /tmp/<descriptive-slug>-task.md. This artifact enables a new session to resume well-scoped execution without relying on conversation memory.
A cold-start task briefing separates execution by structuring the task file into reversible autonomous steps and irreversible actions. The irreversible actions require explicit engineer authorization, ensuring safety gating before any shared-state modifications occur.
Use a task briefing for abandoned PRs or settled tickets needing a cold-start execution artifact. A mid-flight handoff is for continuing active work, while a task briefing resumes well-scoped work without prior conversation memory.
The task briefing artifact requires writing to /tmp/<descriptive-slug>-task.md. This file must include a required preamble and structured sections that separate autonomously safe steps from irreversible actions needing explicit engineer authorization.
No dependencies are required to generate a task briefing for abandoned git workflows. The skill operates independently to produce the cold-start task file in the /tmp directory for your settled tickets or follow-up work.
Task briefing execution gates shared-state actions because the skill enforces a safety-first execution model. It splits reversible autonomous steps from irreversible actions, requiring explicit engineer authorization to prevent unintended modifications.
Nightwolf7570Create a reusable Claude Skill folder with frontmatter and starter instructions.
OmniNode-aiRevoke Edit/Write authorization by deleting a session-specific auth file.
HuuBarExtract session metadata and summaries from Claude Code artifacts into a resume briefing.
TakfesCompacts conversations into redacted handoff documents for the next agent.
