agentsclimarketplace

Esper atom

Skill sichengchen/esper/skills/esper-atom

Tool-neutral, simple workflow layer for agentic, spec-driven software development. 🪄✨

Install
npx -y skills add sichengchen/esper --skill esper-atom

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 1 stars1 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Create or refine a single increment. Smart stage detection — no active creates one, incomplete refines, ready waits for `esper:go`.

SKILL.md

2.6 KB, 574 tokens by cl100k_base, as published. Nobody here has run it

You are working in atom mode — focused on a single increment at a time.

Step 1: Read context

Run esperkit context get to understand the current state. Read .esper/CONSTITUTION.md for project constraints.

Step 2: Smart stage detection

Check the context for active_increment:

No active increment

  1. Ask the user what they want to work on
  2. Run esperkit spec index to show available specs
  3. Help the user define the increment scope:
    • Title (concise, imperative — "Add user authentication")
    • Type: feature, fix, or chore
    • Which spec file is relevant (if any)
    • Which spec section is relevant (if any)
  4. Run esperkit increment create --title "..." --lane atomic --type <type> [--spec <file>] [--spec_section <section>]
  5. Run esperkit increment activate <filename>
  6. Read the created increment file
  7. Help the user fill in the body sections:
    • Context: what exists and why this change matters
    • Scope: what will change (specific, enumerated items)
    • Files Affected: file paths with create/modify annotations
    • Verification: exact commands and expected outcomes
    • Spec Impact: which specs may need updating
  8. Write the refined increment file

Active increment — incomplete as document

If the increment's body sections still have placeholder text:

  1. Read the full increment file
  2. Read referenced spec files
  3. Help the user refine the incomplete sections
  4. Write updates to the increment file

Active increment — fully authored

If all body sections are filled in:

  1. Present a summary of the increment
  2. Tell the user the increment is ready for implementation
  3. Say: "Run esper:go to begin implementation."

Step 3: Plan authoring loop

Stay in the plan-authoring loop:

  1. Present the current increment state
  2. Accept refinements from the user
  3. Write updates
  4. Repeat until the user is satisfied

Do NOT begin implementation. Atom mode is for authoring the increment plan. Implementation starts when the user runs esper:go.

Available CLI commands

  • esperkit context get — read current context
  • esperkit increment list — list increments
  • esperkit increment create --title "..." --lane atomic [--type feature|fix|chore] [--spec <file>] [--spec_section <section>] — create
  • esperkit increment activate <file> — activate
  • esperkit spec index — show spec tree
  • esperkit spec get <file> — read a spec

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 327,069. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.