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/abdelstark/worldforge/persistence-statenpx skills add AbdelStark/worldforge --skill persistence-stategit clone --depth 1 https://github.com/AbdelStark/worldforgeWhat 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 | $0.00082 | $0.00840 |
| Opus 5 | $0.00041 | $0.00420 |
| Sonnet 5 | $0.00016 | $0.00168 |
| Haiku 4.5 | $0.00008 | $0.00084 |
Grade A, and why
persistence-state 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 3d 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 — 58 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Persistence And State
Contract
- WorldForge has no symbolic world store. The mutable
Worldruntime,create_world/save_world/load_world/export_world/import_world/fork_world,.worldforge/worlds, and theworldforge worldCLI were removed in the latent-planning pivot. Planning is done withLatentMPCControllerover thescore/predictcapabilities;worldforge predictrolls a single move through apredictprovider over an in-processworld_statedict (not persisted). - The remaining local state is run workspaces and artifacts under
.worldforge/(defaultstate_dir): run manifests, evidence/issue bundles, evaluation/benchmark reports, and run-index data. It is local-first and single-writer; hosts own durable storage. - All public/semi-public artifacts are JSON-native (string keys, finite numbers, lists, objects,
bools, strings, null). Malformed persisted/provider artifacts raise
WorldStateError; never silently coerce invalid state. - Treat existing
.worldforge/user state as off-limits unless the user points at a disposable directory. Tests must usetmp_path.
Workflow
- Build provider input
world_stateas a plain dict (e.g.{"schema_version":1,"id":..., "scene":{"objects":{obj.id: SceneObject(...).to_dict()}}}); pass it toforge.predict(...). There is noWorldobject to mutate or persist. - Use the run-workspace surface (
worldforge runs ...,runs_prune, evidence bundles) for anything that must survive a process: write through the validated workspace/manifest APIs insrc/worldforge/harness/workspace.pyandsrc/worldforge/runs_prune.py. - Validate serialized artifacts before filesystem writes and after reads; reject unsafe IDs, traversal, and non-JSON-native fields loudly.
- Keep CLI handlers thin; do not duplicate validation or file I/O rules in
src/worldforge/cli.py. - Test flows with
tmp_pathstate dirs. Never use or delete the user's real.worldforge/state.
What ships with it
1 file 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.
- 3d ago First seen · 58 lines · 82 tokens per session scan A 4f352e2d7d5b
persistence-state is a skill published in the GitHub repository AbdelStark/worldforge (108 stars, last pushed 24d ago), licensed MIT. It adds 82 tokens to every session and 840 once invoked, about $0.0004 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 2026-08-30.
Other skills, from other repositories
xie-saining-research-taste
Use when evaluating AI/ML research ideas, choosing research directions, framing problems, planning experiments, improving research taste or paper storytelling, resisting trend-following, or applying Saining Xie's interview-derived views on visual intelligence, world models, research culture, and scientific taste.
improving-frontend-coverage
Runs frontend unit tests with coverage, analyzes coverage reports, and implements meaningful tests to increase coverage by 0.2%. Use when you want to systematically improve frontend test coverage with high-value test cases.
hf-gradio
Use Gradio applications via API. Use when the user asks for to generate a prediction from a Gradio app on Hugging Face spaces or public URL. For example, "Generate an image using black-forest-labs/FLUX.2-dev".
developing-with-streamlit
Use for ALL Streamlit tasks: creating, editing, debugging, beautifying, styling, theming, optimizing, or deploying Streamlit apps. Also custom components, st.components.v2, HTML/JS/CSS work. Discovers and loads version-matched reference docs from the user's installed Streamlit (>=1.57). Triggers: streamlit, st.…
hugging-face-paper-publisher
Publish and manage research papers on Hugging Face Hub. Supports creating paper pages, linking papers to models/datasets, claiming authorship, and generating professional markdown-based research articles.
pr-review
Review a pull request and emit a validated review payload.