From cofoundr
Resume a Claude Code session from the latest epic handoff. Reads docs/handoff/<epic>.md, briefs the agent on epic progress + next-up feature, then waits for "go".
How this skill is triggered — by the user, by Claude, or both
Slash command
/cofoundr:pickupWhen to use
Trigger when the user says "pickup", "/pickup", "resume", "where were we", "pick up where we left off", or "continue from yesterday". Also offer proactively at session start if docs/handoff/ contains any non-archived files. Note: bare /resume is reserved by Claude Code; this is the namespaced equivalent.
The summary Claude sees in its skill listing — used to decide when to auto-load this skill
Read the latest epic handoff and brief the agent on progress + what's next. Wait for "go" before doing any work.
Read the latest epic handoff and brief the agent on progress + what's next. Wait for "go" before doing any work.
One file per epic. Features accumulate inside it. Never overwrites. Archive only when the whole epic is done.
An epic is a coherent body of work — a milestone, a campaign, or a multi-feature push — that spans multiple sessions. The handoff file (docs/handoff/<epic-slug>.md) is the running history of the whole campaign. This skill reads it and briefs the agent so the next session can continue without re-explaining context.
For the section semantics and how the file is written, see the /cofoundr:handoff skill.
/pickup./pickup is a session-start tool. Confirm the user wants to switch context first.List active handoffs. Read filenames from docs/handoff/*.md (excluding docs/handoff/archived/). Sort by Last updated desc.
If exactly one active handoff: brief on that one. If zero active handoffs: say "No active epics found. Want to start a new one?" and stop. If multiple active handoffs: show a short list and ask which:
Active epics:
1. funnel-rewire (last updated 2 days ago, session #6, 3 features shipped)
2. billing-v2 (last updated 5 days ago, session #2, 0 features shipped)
3. auth-migration (last updated 14 days ago, session #1) ⚠ stale
Which epic?
Read the chosen handoff file. Then read the project spec source of truth in this order: agents.md (CoFoundr-system + AGENTS.md convention) → docs/spec.md → CLAUDE.md → README.md.
Output the "You are here" briefing in this exact shape:
## You are here
Project: <name from spec>
Epic: <slug from filename>
Status: <from handoff>
Sessions so far: <count of session log entries>
Features shipped: <n> · Features remaining: <m>
Last session built: <one-sentence from "Current state">
Recently shipped:
- <most recent 2-3 from "Features completed">
Open questions (need your input):
- <each open question>
(or "_(none)_" if empty)
Top tech debt:
- [<sev>] <issue>
(top 3 by severity)
Next priority: <#1 from "Next 3 priorities">
Success criterion: <its success criterion>
Ready to continue? Reply "go" to start, or tell me to do something else.
Status: blocked or any open questions, surface that before the next-up priority.docs/handoff/ directory exists yet. Say "No handoffs found. Want to start an epic with /cofoundr:handoff?" and stop.Status: archived (file is in docs/handoff/archived/, not in active list). Don't brief on it. If the user explicitly asks for an archived epic, tell them when it was archived and ask if they want to re-open it (move back to docs/handoff/).Provides behavioral guidelines to reduce common LLM coding mistakes, focusing on simplicity, surgical changes, assumption surfacing, and verifiable success criteria.
Searches, retrieves, and installs Agent Skills from prompts.chat registry using MCP tools like search_skills and get_skill. Activates for finding skills, browsing catalogs, or extending Claude.
Creates, edits, and optimizes skills for Claude Code, including drafting, evaluating with test prompts, iterating on performance, and improving skill descriptions for better triggering accuracy.
npx claudepluginhub cofoundr-org/handoff-skill