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 skills add joacod/skills --skill project-specsgit clone --depth 1 https://github.com/joacod/skillsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/joacod/skills/project-specs)<a href="https://agentmods.dev/skills/joacod/skills/project-specs"><img src="https://agentmods.dev/badge/skills/joacod/skills/project-specs/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/joacod/skills/project-specs"><img src="https://agentmods.dev/badge/skills/joacod/skills/project-specs.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What 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 5.1 | $0.00191 | $0.01269 |
| Opus 5.5 | $0.00076 | $0.00508 |
| Sonnet 5.5 | $0.00038 | $0.00254 |
| Haiku 4.5 | $0.00019 | $0.00127 |
Grade A, and why
project-specs scanned grade A with 0 findings 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 1mo ago.
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.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 118 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Project Specs
Capture a finished exploration as a durable, living implementation package inside the repository. The package must hold the desired last form of the product, the full version roadmap toward that destination, and enough context that an agent who never saw the brainstorm can implement the current slice and later continue after the spec is updated.
This skill is the step immediately before implementation, and the spec refresh after each meaningful step:
brainstorm → specification → version roadmap → executable near-term tasks → implement → update spec → next version
It does not implement the project.
Choose the mode
- Create when no project specification package exists yet. Write the destination, the whole roadmap, and executable tasks for the first version.
- Update when a matching package already exists, including after a version, task, or experiment landed. Read it first and follow the update behavior in spec-structure.md. Implementation reality wins over outdated planning assumptions.
If the request is only a small coding task, an ordinary design question, a main README rewrite, or general Markdown cleanup, do not use this skill.
Operate in this order
Follow this pipeline; do not begin by filling a template:
gather → inspect → classify → research → specify → decompose → review → summarize → stop
- Gather. Use whatever is available: the current conversation, user notes or handoffs, existing code and experiments, mentioned URLs, and previous spec files. Do not require every input. Do not ask for missing optional material unless two interpretations would produce materially different packages.
- Inspect. Map the repository enough to reuse real infrastructure: stack, packages, conventions, existing experiments, constraints, and nearby code. Prefer concrete paths over generic architecture. Do not read the entire codebase blindly.
- Classify. Separate what is decided, provisional, speculative, or in contradiction with earlier brainstorm turns. Identify assumptions that need experiments, technical risks, existing systems to reuse, things that should not be built yet, and premature abstractions.
- Research. Preserve every important external reference. If browsing is available and understanding a linked repository, article, demo, paper, model, or tool would materially improve the spec, inspect it instead of judging it from its title. Catalog what to borrow and what not to copy blindly.
- Specify. Read spec-structure.md. Create or
update the project directory. Prefer an established repository documentation
convention when one already exists; otherwise use
docs/<project-slug>/. - Decompose. Read task-design.md. Write the full version sequence through the desired last form. Thoroughly task-split the current near-term version (V0 on create; the next incomplete version on update). Keep far-future versions directional, but present.
- Review. Read quality-checklist.md and fix inconsistencies before finishing.
- Summarize, then stop. Return the compact handoff below. Do not start V0, scaffold production architecture, or "just implement the first task."
What ships with it
5 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 1mo ago First seen · 118 lines · 191 tokens per session scan A d2c0eb795025
project-specs is a skill published in the GitHub repository joacod/skills (3 stars, last pushed 24d ago), licensed MIT. It adds 191 tokens to every session and 1,269 once invoked, about $0.0008 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-10.
Other skills, from other repositories
antfu-issue-triage
Triage open issues and PRs of the current GitHub repo - find already-fixed, duplicate, and spam items, suggest labels, gauge demand from reactions and comments, then present a table of proposed actions and comments for confirmation before touching anything. Use when asked to triage, organize, clean up, or groom issues…
Freelance Empire
Complete freelance business operating system — from first client to six-figure solo business. Covers positioning, pricing, proposals, client management, finances, scaling, and the transition from side-hustle to full-time. Use when starting freelancing, raising rates, managing clients, building recurring revenue, or…
AI Recruiting Engine
Full-cycle recruiting agent — source, screen, score, and hire top talent with structured frameworks, scorecards, and pipeline automation. Zero dependencies.
agent-council
Complete toolkit for creating autonomous AI agents and managing Discord channels for OpenClaw. Use when setting up multi-agent systems, creating new agents, or managing Discord channel organization.
agent-swarm-workflow
Jeffrey Emanuel's multi-agent implementation workflow using NTM, Agent Mail, Beads, and BV. The execution phase that follows planning and bead creation. Includes exact prompts used.
agi-farm
Interactive setup wizard that creates a fully working multi-agent AI team on OpenClaw. One command bootstraps agents, SOUL.md personas, comms infrastructure (inboxes/outboxes/broadcast), cron jobs, auto-dispatcher (HITL + rate-limit backoff + dependency checking), and a portable GitHub bundle — all customized to team…