Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add skills/loopdive/js2/tech-lead-loopnpx skills add loopdive/js2 --skill tech-lead-loopgit clone --depth 1 https://github.com/loopdive/js2What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00055 | $0.03861 |
| Opus 5 | $0.00028 | $0.01930 |
| Sonnet 5 | $0.00011 | $0.00772 |
| Haiku 4.5 | $0.00006 | $0.00386 |
Grade B, and why
tech-lead-loop scanned grade B with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Reads agent configuration directoriesmediumAgent snooping
.claude/, .codex/, .gemini/ hold keys, settings and other credentials a mod has no legitimate need for.
cat .claude/ci-status/pr-$p.json 2>/dev/null | \ How it starts
The opening of the file, as written. The whole thing — 318 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Tech Lead Loop
The tech lead is the orchestrator of a multi-dev sprint. This skill defines the standing loop the tech lead runs to keep development unblocked, enforce protocol, and move the merge queue forward — without becoming a bottleneck.
The loop has five phases. You cycle through them continuously for the duration of the sprint. Each phase has triggers, a checklist, and the specific sub-skills / slash commands to invoke.
Never delegate understanding. This skill tells you what to check; you must read the results and decide.
Principles
- Devs are authoritative for their own PRs. Tech lead only intervenes when a dev is truly stuck, when claims conflict, or when the merge queue stalls.
- TaskList is the canonical dispatch mechanism. Populate it at sprint start and whenever new issues are added mid-sprint. Devs self-serve — they claim tasks, self-merge, and claim the next task without waiting for tech lead direction.
- SendMessage only for escalations and exceptions. Devs contact tech lead only if TaskList is empty, they are blocked >30 min, or CI criteria force an escalation. Tech lead contacts devs only when they are stuck/silent or there is a file conflict to resolve.
- Every protocol correction issued to one dev should be generalized. If you correct one dev, broadcast the canonical protocol to all devs within the next cycle — individual corrections rot.
- Keep tool calls parallel when independent. One message with five Bash calls is dramatically faster than five sequential ones.
- Measure the merge queue, not the task queue. The right metric is "how many PRs are ready to merge but unmerged" — stalls there compound.
Phase 0 — Session start / context load
Trigger: fresh conversation, or resume after /compact or context rollover.
Checklist:
- Read the current sprint doc:
plan/issues/sprints/<N>/sprint.md - Read the tech lead handoff:
plan/agent-context/tech-lead.md git status && git branch --show-current && git log --oneline origin/main -10gh pr list --state open --limit 20— snapshot of inherited open PRsls .claude/worktrees/— inherited worktrees (may contain WIP)- Read
plan/method/file-locks.mdif it exists — active dev claims - Read
.claude/memory/MEMORY.mdtop section — critical rules + feedback index TaskList— if empty, create tasks fromplan/issues/sprints/{N}/(current sprint dir, filter bystatus: ready) before touching any dev.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- yesterday First seen · 318 lines · 55 tokens per session scan B 8a4e46928621
tech-lead-loop is a skill published in the GitHub repository loopdive/js2 (59 stars, last pushed yesterday), licensed Apache-2.0. It adds 55 tokens to every session and 3,861 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it B with 1 finding (reads agent configuration directories). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other skills, from other repositories
migrate-oxfmt
Guide for migrating a project from Prettier or Biome to Oxfmt. Use when asked to migrate, convert, or switch a JavaScript/TypeScript project's formatter from Prettier or Biome to Oxfmt.
migrate-oxlint
Guide for migrating a project from ESLint to Oxlint. Use when asked to migrate, convert, or switch a JavaScript/TypeScript project's linter from ESLint to Oxlint.
insta-snapshots
Guide for working with and updating insta snapshot tests in Oxc without terminal interaction.
performance-lint-rules
Performance review guidance for Oxc linter rule implementations. Use only when reviewing Rust rule code under crates/oxclinter/src/rules/ or when explicitly auditing those rules for performance improvements.
oxc-docs
Comprehensive reference for the JavaScript Oxidation Compiler (Oxc) — a collection of high-performance JavaScript tools written in Rust. Covers parser design (lexer, AST, parser, errors, semantic analysis), architecture (parser, linter, test infrastructure, AST tools), ECMAScript specification and grammar, performance…
babel-docs
Babel 7.x/8.x — plugins, presets, config, @babel/core, parser, traverse, types, CLI, preset-env.