Borrowing it
Nothing to install: this file belongs to tatargabor/set-core. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/tatargabor/set-core/main/.claude/skills/set/glm/SKILL.mdgit clone --depth 1 https://github.com/tatargabor/set-coreWrote 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/tatargabor/set-core/glm)<a href="https://agentmods.dev/skills/tatargabor/set-core/glm"><img src="https://agentmods.dev/badge/skills/tatargabor/set-core/glm.svg" alt="Measured on agentmods" 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 | $0.00115 | $0.02981 |
| Opus 5 | $0.00057 | $0.01491 |
| Sonnet 5 | $0.00023 | $0.00596 |
| Haiku 4.5 | $0.00012 | $0.00298 |
Grade A, and why
glm 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 today.
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 — 199 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Running GLM — set-glm
The hard part is not switching the model — it is the context window. The Claude Code CLI does not know GLM's window, so it cuts to a conservative value (~200k) — and a long prompt either fails loudly or runs in a silent compact loop. This skill exists so that nobody has to measure that again.
⚠ Switching is always a USER decision. No automatics and no silent fallback: on missing config the runner stops, it does not fall back to Claude. A silent fallback is the worst outcome because the run would still finish — just inside the other framework, with nothing telling you which one.
Usage
set-glm --check # config + LIVE probe call — start here
set-glm -p "prompt" # one-shot call
set-glm -p "…" --output-format json # structured output (with token- and cost-fields)
set-glm # interactive session on GLM
set-glm --print-env # what it sets (token masked)
Every other flag passes through to claude unchanged (--allowedTools,
--json-schema, --append-system-prompt, …). --model and --autocompact are only
added when the caller did not supply them.
Configuration — ONE central file (since 2026-08-29)
~/.config/set-core/providers.json, mode 0600. This is the only place the framework
reads a provider credential from; set-glm and the fleet's agent-owner call the same
resolver, so the measured launch parameters live in one place and do not drift apart.
set-providers path # where it is expected
set-providers show # what is in it — token masked
set-providers migrate # carry the old glm.env over, ONE command
Precedence has three levels: machine-level default → project override → the request itself. The model is decided per field; the credential and its endpoint are ONE block — a level either provides both or is rejected. A key is issued for one endpoint, and taking the two from different levels is a combination nobody has described: best case a 401, worst case the other bill.
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.
- today First seen · 199 lines · 115 tokens per session scan A 7e0c099052f4
glm is a skill published in the GitHub repository tatargabor/set-core (35 stars, last pushed today), licensed MIT. It adds 115 tokens to every session and 2,981 once invoked, about $0.0006 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 2026-09-06.
Other skills, from other repositories
ai-engineer
Expert knowledge in AI/ML development, model deployment, and MLOps practices.
prompt-craft
Translate natural language image descriptions into detailed, structured DALL-E prompts with subject, style, composition, lighting, and mood specifications.
hatch3r-cli-jq
JSON processor and query language. Use when shaping JSON streams via jq-syntax filters and select expressions; invoke jq. Reads stdin and emits stdout; integrates seamlessly into shell pipelines.
agent-v3-memory-specialist
Agent skill for v3-memory-specialist - invoke with $agent-v3-memory-specialist.
agent-neural-network
Agent skill for neural-network - invoke with $agent-neural-network.
embeddings
Vector embeddings with HNSW indexing, sql.js persistence, and hyperbolic support. 75x faster with agentic-flow integration. Use when: semantic search, pattern matching, similarity queries, knowledge retrieval. Skip when: exact text matching, simple lookups, no semantic understanding needed.