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/albertbuchard/forge/forge-codexnpx skills add albertbuchard/forge --skill forge-codexgit clone --depth 1 https://github.com/albertbuchard/forgeWhat 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 | $0.00111 | $0.16064 |
| Opus 5 | $0.00056 | $0.08032 |
| Sonnet 5 | $0.00022 | $0.03213 |
| Haiku 4.5 | $0.00011 | $0.01606 |
Grade A, and why
forge-codex 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 2d 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 — 1,014 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Forge Codex
Use this plugin when you want Codex to work directly with Forge through the curated MCP tool surface.
Live Contract And Missing-Information Gate
Before the first Forge read or write in a session, call
forge_get_agent_onboarding. Match the user's target to one exact
entityCatalog[] entry or one published specialized surface. Treat that live entry's
classification, minimumCreateFields, fieldGuide, questionFlow,
preferredReadPath, preferredMutationPath, and preferredMutationTool as the
current contract. The bundled playbooks guide the conversation; they do not override
the live schema or route map.
Build a private missing-information diff before asking anything:
- remove details the user already supplied
- remove optional fields and published defaults that do not change meaning, accountability, timing, retrieval, safety, or route selection
- on update, read the current record first and preserve every field the user did not ask to change
- ask one Psyche question at a time; for logistical records, one compact question may group inseparable details such as start, end, and timezone
- act when no blocking ambiguity remains instead of asking a polished extra question
If the target is absent from live onboarding, refresh once. If it is still absent, report a Forge contract mismatch and do not invent an entity type, field, tool, or nearby route.
Forge has planning, health, preferences, Psyche, questionnaire, self-observation,
wiki surfaces, the Artifact Store, read-model surfaces, and specialized Movement, Life Events, Life Force, and Workbench domain surfaces.
The planning side covers goals, projects, strategies, tasks,
habits, tags, notes, calendar events, recurring work blocks, task timeboxes, live
task runs, and agent-authored insights. The health side covers sleep_session,
workout_session, and the read-only training_load surface for cardiovascular
load and HR zone review, plus the weight_loss read and nutrition evidence
workflow. The preferences side covers preference_catalog,
preference_catalog_item, preference_context, and preference_item plus the game,
judgments, and signals. The Psyche side covers values, patterns, behaviors, beliefs,
modes, guided mode sessions, flashcards, trigger reports, event types, reusable emotion
definitions, questionnaire_instrument, questionnaire_run, and the note-backed
self-observation calendar. The Artifact Store is a specialized CRUD surface for
trusted stored files such as spreadsheets, documents, PDFs, text, structured text,
and images; artifact relationships use the general entity_links model, and Codex
must not download, decrypt, open, execute, preview, transform stored file bytes, or
submit artifact passwords. Read-model surfaces include operator_overview,
operator_context, calendar_overview, sleep_overview, sports_overview,
training_load, weight_loss, and the self-observation calendar; ask what
practical decision the read should support before adding write-shaped questions.
Preferences Workspace is also read-model-only. Use it to explain inferred scores from
judgments, signals, overrides, evidence count, and uncertainty before offering a
dedicated Preferences action. A workspace read never initializes or refreshes state;
if it is missing, report that state and use an explicit Preferences action only after
the user chooses it.
Movement, Life Events, Life Force, and Workbench use dedicated route
families and must not be forced through batch CRUD. Forge is explicitly multi-user: every stored entity can
belong to a typed human or bot user through userId, reads can scope to one or
many users with userId or repeated userIds, and cross-user links are valid when
the request is intentional.
What ships with it
2 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.
- 2d ago First seen · 1,014 lines · 111 tokens per session scan A 7a66357168c9
forge-codex is a skill published in the GitHub repository albertbuchard/forge (23 stars, last pushed 5d ago), licensed Apache-2.0. It adds 111 tokens to every session and 16,064 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 2026-08-30.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
brainstorming
You MUST use this before any creative work - creating features, building components, adding functionality, or modifying behavior. Explores user intent, requirements and design before implementation.
chat-pet-sprite-creation
Use when creating or changing VS Code chat pet sprite art, sprite sheets, state animations, eye treatments, Stable/Insiders variants, or pet transitions under src/vs/workbench/contrib/chat/browser/widget/media/chatPet.
cpu-profile-analysis
Analyze V8/Chrome CPU profiles (.cpuprofile) and DevTools trace files (Trace-.json). Use when: profiling performance, investigating slow functions, comparing code paths, finding bottlenecks, analyzing timeToRequest, understanding call trees from sampling profiler data, analyzing layout/paint/rendering, investigating…
agent-host-chat-contributions
Build and review cross-cutting agent-host chat behavior through lifecycle contributions. Use when adding turn lifecycle side effects, prompt or context injection, restored-history transformation, protocol-action observation, or when reviewing changes that add code to AgentSideEffects or AgentService.
auto-perf-optimize
Run agent-driven VS Code performance or memory investigations. Use when asked to launch Code OSS, automate a VS Code scenario, run the Chat memory smoke runner, capture renderer heap snapshots, take workflow screenshots, compare run summaries, or drive a repeatable scenario before heap-snapshot analysis.