resume

Resumes an approved project plan from current state with swarm identity reconciliation.

Updated May 31, 2026
One-click install
npx skills add https://github.com/AlexanderNarbaev/agi --skill resume-alexandernarbaev
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: resume
Source: https://github.com/AlexanderNarbaev/agi/tree/main/.opencode/skills/resume
Command: npx skills add https://github.com/AlexanderNarbaev/agi --skill resume-alexandernarbaev

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Continuing work on an existing project risks acting on stale plans, mismatched swarm identities, and outdated context left by previous sessions. This protocol safely resumes an approved plan by validating state before execution. ## Core Features & Use Cases - Plan State Detection: Checks for .swarm/plan.md and reads its Swarm field to decide whether to resume or start a new project. - Swarm Identity Reconciliation: When the plan's swarm id differs from the active one, it updates plan.md and context.md, purges stale memory blocks, and clears the SME Cache section. - New Project Bootstrap: Runs complexity_hotspots for a 90-day risk map, optionally extracts technical debt with todo_extract, and captures a baseline SBOM via sbom_generate. - Use Case: A developer returns to a project previously managed by a different swarm; the protocol clears stale context, aligns identity fields, and resumes at the current task without manual cleanup. ## Quick Start Resume my existing project plan from its current task and clear any stale context from the previous swarm.

Frequently Asked Questions about resume

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I resume an existing project plan in a swarm workflow?

The protocol checks for .swarm/plan.md and reads its Swarm field. If the field matches the active swarm id, it resumes at the current task; otherwise it reconciles identity fields before continuing.

What happens when the plan's swarm id differs from the active swarm?

The plan.md and context.md Swarm fields are updated to the active swarm id, memory blocks referencing the old identity are purged, the SME Cache section is deleted, and the user is informed before resuming.

What happens if no plan file exists when resuming?

If .swarm/plan.md does not exist, the project is treated as new and the protocol proceeds to clarification mode, then runs complexity_hotspots, optionally todo_extract, and sbom_generate for a baseline inventory.

How does the protocol handle risk assessment for new projects?

It runs the complexity_hotspots tool over 90 days to build a risk map, noting modules flagged security_review or full_gates in context.md so stricter QA gates apply during Phase 5.

What are the limitations of the resume protocol?

It depends entirely on the presence and correctness of .swarm/plan.md and context.md. If these files are corrupted or missing key fields, identity reconciliation and task resumption cannot proceed reliably.