develop

Coordinates implementation work from goal clarification through planning, execution, and verification.

1|Updated Mar 6, 2026
One-click install
npx skills add https://github.com/peregrinus879/eyragents --skill develop-peregrinus879
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: develop
Source: https://github.com/peregrinus879/eyragents/tree/main/agents/.agents/skills/develop
Command: npx skills add https://github.com/peregrinus879/eyragents --skill develop-peregrinus879

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Complex development work often loses track of scope, decisions, and verification evidence across long sessions, context compaction, or handoffs. This Skill provides a disciplined workflow that carries substantive work from goal clarification through planning, execution, and verification while preserving resumable state. ## Core Features & Use Cases - Workstream State Management: Maintains a private gitignored checkpoint under .eyr-plans/ capturing outcome, plan, decisions, evidence, and next actions so work survives pauses, compaction, and resumption. - Atomic Change Planning: Records independently valid atomic changes with repository/file scope, ownership, dependencies, and verification before substantial implementation or delegation. - Verification Contract: Defines acceptance checks during planning, runs repository gates like make lint, make check, and make verify, and requires evidence-based reuse rules before declaring completion. - Use Case: You ask the agent to implement a multi-file feature across two repositories. The skill plans atomic commits, saves checkpoints before long operations, runs the prescribed gates, and finishes with a clear next-step choice to hand verified changes to the commit workflow. ## Quick Start Ask the agent to develop and verify a specific feature or fix, and it will plan atomic changes, execute them, run the repository checks, and preserve resumable state.

Frequently Asked Questions about develop

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

FAQPage Schema
How do I resume interrupted AI coding work after context compaction?

The skill stores a checkpoint under a gitignored `.eyr-plans/<workstream>/` directory capturing outcome, plan, decisions, evidence, and next action. On resume, the agent re-reads the checkpoint and actual Git state, reconciles drift, and continues at the current step.

How to plan atomic commits before implementing a large feature?

Before substantial implementation, the skill records independently valid atomic changes with repository or hunk scope, ownership, dependencies, and verification. Coupled code, configuration, tests, and documentation stay together, and the map updates as findings change.

What verification checks run before declaring implementation complete?

The verification contract runs the repository's prescribed gates such as `make lint`, `make check`, `make restow`, and `make verify` when defined. Evidence is recorded with actual commands and results, and failed, skipped, or unverified checks are reported explicitly.

Can prior test results be reused instead of rerunning checks?

Reuse is allowed only after re-reading the original passing evidence and confirming the exact tested contents and relevant inputs are unchanged. Unknown inputs, untracked contents, or source changes forbid reuse, and commit preparation must validate the exact staged candidate.

When does the skill hand work off to a commit workflow?

After verified implementation with intended uncommitted changes, the skill presents a next-step selector offering to prepare commit reviews or pause. A clear request to commit or finish the Git workflow proceeds directly to the commit skill without another question.