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 lossless-group/lossless-agent-skills --skill study-repos-firstgit clone --depth 1 https://github.com/lossless-group/lossless-agent-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/lossless-group/lossless-agent-skills/study-repos-first)<a href="https://agentmods.dev/skills/lossless-group/lossless-agent-skills/study-repos-first"><img src="https://agentmods.dev/badge/skills/lossless-group/lossless-agent-skills/study-repos-first/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/lossless-group/lossless-agent-skills/study-repos-first"><img src="https://agentmods.dev/badge/skills/lossless-group/lossless-agent-skills/study-repos-first.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 | $0.00158 | $0.02283 |
| Opus 5 | $0.00079 | $0.01141 |
| Sonnet 5 | $0.00032 | $0.00457 |
| Haiku 4.5 | $0.00016 | $0.00228 |
Grade A, and why
study-repos-first 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 12d 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 — 123 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Study Repos First
A study is a curated reference collection — a directory of upstream specs, prior art, reference implementations, papers, or codebases — pinned as git submodules before a decision is made or a feature is built. It is not a project (it ships nothing) and not documentation (it lives, it's checked-out code). It is a reading list with structure, pinned so it never rots out from under us.
This skill is foundational and iterative. It will be expanded as we work through real studies together. Treat the contents as the current best understanding, not the final word.
When to use this skill
- Starting any task in
ai-labs/(or any pseudomonorepo) where the question is about conventions — naming, file shapes, folder layouts, schemas, protocols, manifest formats — rather than novel logic - The user says "let's study X", "what's the prior art on X", "how do other projects handle X", "pin a reference", "add a submodule under studies/"
- A design decision is on the table and the temptation is to let the agent fill in patterns from training data
- Scaffolding a new
studies/<topic-slug>/directory - Adding a reference to an existing study (new submodule, README update)
- Deciding whether a study has outgrown a subdirectory and should be promoted to its own repo (the
lossless-group/study-<slug>pattern) - Reading from a study to ground a current implementation question
The behavioral core (this is the actual skill)
Don't paraphrase from training data when pinned upstream code is — or could be — one cd away.
When a domain question comes up:
- Check if a study already covers it. Look in
ai-labs/studies/(and anystudies/directory walking up the tree). If a study exists, read from the pinned submodules, not from memory. - If no study exists and the question is meaningful, propose one. Frame the question first; pick a slug; set up the directory.
- Cite findings with file paths —
studies/<topic>/<repo>/<file>:<line>— so the user can jump to the upstream source. - Do not summarize or paraphrase upstream specs into prose unless asked. The upstream repo is the artifact. A paragraph paraphrasing
llms.txtis a hot take that will rot.
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.
- 12d ago First seen · 123 lines · 158 tokens per session scan A d8d0944935ed
study-repos-first is a skill published in the GitHub repository lossless-group/lossless-agent-skills (4 stars, last pushed 19d ago), licensed MIT. It adds 158 tokens to every session and 2,283 once invoked, about $0.0008 per session on Opus 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 2026-08-31.
Other skills, from other repositories
session-review
End-of-session adversarial review loop. Assemble the session's work into a role-assigned, self-contained brief, then run independent reviewers in parallel — an isolated code-reader (the idea-validator agent) that reads the ACTUAL files and web-checks technology currency, plus an external-family model if you have one …
memory-audit
Audit MEMORY.md against the memory discipline — oversized sections, settled multi-session patterns that belong in knowledge/concepts/, stacked chronicle blocks, stale entries. Produces a move plan as a table for approval, then executes the approved moves atomically. Use when the SessionStart hook reports a tripped…
second-opinion
Cross-check the agent's own answer with independent reviewers before bringing it to the user. Use when the user says 'second opinion', 'sanity check', 'cross-check', 'am I missing something', 'stress-test', 'devil's advocate', 'run a full review', 'this is important', 'high-stakes', 'help me choose between', 'critique…
close-session
An end-of-session workflow that reviews new lessons, asks before promoting repeated patterns, refreshes memory files, and writes a handoff.
next-cache-components
Next.js 16 Cache Components guidance. Use when refactoring React Server Components for performance, debugging Partial Prerendering (PPR) issues, or applying the use cache directive, cacheLife, cacheTag, updateTag, and revalidateTag. Also use when deciding whether data should be static, cached, or dynamic, or when…
agents-sdk
Build AI agents on Cloudflare Workers using the Agents SDK or build Model Context Protocol (MCP) servers. Load when creating stateful agents, durable workflows, MCP servers, checking MCP schema/error responses, or reviewing MCP code quality. Covers Agent class, state management, callable RPC, Workflows integration…