Now liveThe Skillselion MCP - thousands of ranked skills, loaded into your agent mid-task. No install.Get it →
othmanadi avatar

Apex Tier3

  • 4 installs
  • 2 repo stars
  • Updated March 14, 2026
  • othmanadi/apex

Pair-program in real time with a propose-confirm loop and decision log for high-stakes architecture, debugging, or exploratory work.

About

Runs a real-time pair-programming mode where the agent proposes and the human confirms every significant decision. A developer uses it for high-stakes architecture design, critical debugging, or exploratory work needing discussion at each step.

  • Propose-confirm loop: never write >50 lines without checking in
  • Maintains a running decision log with rationale and alternatives

Apex Tier3 by the numbers

  • 4 all-time installs (skills.sh)
  • +1 installs in the week ending Aug 4, 2026 (Skillselion tracking)
  • Ranked #2,331 of 3,282 Productivity & Planning skills by installs in the Skillselion catalog
  • Data as of Aug 4, 2026 (Skillselion catalog sync)
npx skills add https://github.com/othmanadi/apex --skill apex-tier3

Add your badge

Show developers this skill is listed on Skillselion. Paste this into your README.

Listed on Skillselion
Installs4
repo stars2
Last updatedMarch 14, 2026
Repositoryothmanadi/apex

What it does

Pair-program in real time with a propose-confirm loop and decision log for high-stakes architecture, debugging, or exploratory work.

Files

SKILL.mdMarkdownGitHub ↗

APEX Tier 3 — Pair Mode

Work side-by-side with the human in real-time. Every significant decision is discussed. The agent proposes, the human approves or redirects.

When to Use

Use Tier 3 when:

  • Designing new architecture from scratch
  • Debugging a critical production issue
  • The problem space is ambiguous or exploratory
  • The human wants to learn from the process
  • Stakes are high and mistakes are expensive

Interaction Protocol

The Propose-Confirm Loop

Every action follows this pattern:

1. Agent proposes — "I think we should do X because Y. Here's what that looks like..." 2. Human confirms, modifies, or rejects — "Yes" / "Do X but change Z" / "No, try A instead" 3. Agent executes — Implements the confirmed approach 4. Agent reports — "Done. Here's what changed. Ready for next step."

Rules of Engagement

  • Never write more than 50 lines without checking in
  • Always explain WHY before proposing WHAT
  • Show alternatives when there's a genuine trade-off: "Option A gives us X but costs Y. Option B gives us Z but costs W."
  • Admit uncertainty — "I'm not confident about this approach because..." is always better than guessing
  • No yes-man behavior — If the human's suggestion has a flaw, say so respectfully with evidence

Workflow

Opening

Start by understanding the problem space:

Before we start, I need to understand:
1. What are we trying to achieve?
2. What constraints exist (time, tech, compatibility)?
3. What have you already tried or considered?
4. What does success look like?

During the Session

Maintain a running decision log:

## Decision Log

| # | Decision | Rationale | Alternatives Considered |
|---|----------|-----------|------------------------|
| 1 | Use Prisma over Drizzle | Team familiarity | Drizzle (faster), raw SQL (flexible) |
| 2 | REST over GraphQL | Simpler for this scope | GraphQL (flexible queries) |

After every significant block of work, run validation:

Linux/Mac:

bash "${CLAUDE_SKILL_DIR}/scripts/validate.sh" <target-dir>

Windows:

powershell -File "${CLAUDE_SKILL_DIR}/scripts/validate.ps1" <target-dir>

Closing

At the end of the session:

1. Summarize all decisions made 2. List any open questions or TODOs 3. Identify learnings that should be added to AGENTS.md 4. Commit work with a descriptive message referencing the decision log

Output

Working code with a complete decision log. Any recurring patterns or preferences discovered during the session are candidates for AGENTS.md updates via apex-learn.

Example decision log:

## Decision Log — Event System Architecture

| # | Decision | Rationale | Alternatives Considered |
|---|----------|-----------|------------------------|
| 1 | Use EventEmitter over message queue | Simpler for current scale, can migrate later | RabbitMQ, Redis pub/sub |
| 2 | Typed events with Zod schemas | Runtime validation + TypeScript inference | io-ts, manual types |
| 3 | Async handlers by default | Non-blocking, better throughput | Sync handlers |
| 4 | Dead letter queue for failures | Debugging + replay capability | Log and drop |

## Open Questions
- [ ] Should we add event versioning now or later?
- [ ] Rate limiting for high-frequency events?

## Learnings for AGENTS.md
- Always use typed events (add to Architecture Rules)
- Prefer async handlers unless order matters (add to Preferences)

Related skills

This week in AI coding

Five minutes, every Monday - the tools, releases and tactics for developers.

unsubscribe anytime.