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/corridortech/posecap/ad-domainnpx skills add CorridorTech/PoseCap --skill ad-domaingit clone --depth 1 https://github.com/CorridorTech/PoseCapWhat 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 | $0.00143 | $0.01695 |
| Opus 5 | $0.00072 | $0.00847 |
| Sonnet 5 | $0.00029 | $0.00339 |
| Haiku 4.5 | $0.00014 | $0.00169 |
Grade A, and why
ad-domain 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 — 136 lines — stays where its author put it; the contents beside it link to each section on GitHub.
<background_information>
Implements ADR-0019 — Layer 2 of the artifact stack. Lazy lifecycle owner of CONTEXT.md. Process scaffold for capturing the project's ubiquitous language.
The skill owns capture of vocabulary. The trigger for most updates lives in adjacent skills:
ad-grill-meresolves a term during interview → routes here.ad-specintroduces a new noun while drafting a spec → routes here.ad-architecturenames a domain-bound concept → routes here.ad-driftdetects code/glossary drift → routes here.- Direct
/ad-domaininvocation when the user wants to add a term explicitly.
Codex auto-trigger on description keywords is less mature than Claude Code's. If auto-invocation does not fire when the user mentions a domain term, ubiquitous language, glossary, or naming drift, invoke this skill manually. </background_information>
- Read
CONTEXT.mdat the repo root. If present → single-context repo; updates land here. - Read
CONTEXT-MAP.mdat the repo root. If present → multi-context repo; updates land in the per-contextCONTEXT.mdnamed by the map. - Neither present → no
CONTEXT.mdexists yet. Decide:- Single-context (default): plan to create
CONTEXT.mdat the repo root. - Multi-context: only declare this when the user has explicitly named ≥2 bounded contexts. Plan to create
CONTEXT-MAP.mdplus the relevant per-context file.
- Single-context (default): plan to create
Lazy-creation rule: do not create the file just because the skill was invoked. Create it only when there is at least one resolved term to write.
Step 1 — resolve the term. A glossary entry has four parts:
### <Canonical Noun>
**Definition:** <one-sentence, project-specific, no general programming words>.
_Avoid_: <Alias 1>, <Alias 2>. <Why each is misleading>.
**Related code:** [`path/to/file.ts:42`](path/to/file.ts:42).
Resolution discipline:
- Be opinionated. One canonical name. Others become
_Avoid_aliases with a one-line reason. - One-sentence definitions. If it takes a paragraph, the term is not yet sharp.
- Project-specific only. "Customer", "Triage role", "Materialization cascade" qualify. "Service", "Handler", "Controller" do not (those belong to architectural vocabulary per ADR-0020, not domain).
- Flag conflicts. If two terms could mean the same thing, add the conflict to the Flagged ambiguities section instead of forcing a premature resolution.
Step 2 — locate the insertion point. CONTEXT.md has three canonical sections:
# <Project Name> — Domain Glossary
## Language
<term entries here, one per `### <Term>` heading>
## Relationships
- A **Customer** has many **Orders**.
- An **Order** belongs to one **Customer** and contains one or more **LineItems**.
## Flagged ambiguities
- "Account" vs "User" — see Issue #42; treated interchangeably in code today, separation pending ADR-NNNN.
Insert the new term alphabetically under ## Language. If a relationship clause is implied, append to ## Relationships. If the resolution surfaces an ambiguity rather than closing one, append to ## Flagged ambiguities and mark the term entry "(resolution pending)".
Step 3 — write or update. First creation uses Write; subsequent updates use Edit. Preserve existing entries — never reorder or rewrite past terms while adding new ones.
For first creation, the file shape is:
# <Project Name> — Domain Glossary
_Lazy artifact — only contains terms that have been resolved through grilling, spec drafting, or explicit capture. Empty entries are worse than no entry; speculation belongs elsewhere._
_Maintained by `/ad-domain`._
## Language
### <First resolved term>
...
## Relationships
(empty until the second term resolves)
## Flagged ambiguities
(empty)
For multi-context repos, CONTEXT-MAP.md is the index:
# Context Map
This repository contains multiple bounded contexts. Each has its own `CONTEXT.md`.
- **Ordering** — `src/ordering/CONTEXT.md`. Owns: Customer, Order, LineItem, Cancellation.
- **Billing** — `src/billing/CONTEXT.md`. Owns: Invoice, Payment, Refund.
System-wide ADRs live at `doc/adr/`. Context-scoped ADRs live under each context.
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.
- 2d ago First seen · 136 lines · 143 tokens per session scan A c05e0243f83c
ad-domain is a skill published in the GitHub repository CorridorTech/PoseCap (190 stars, last pushed 10d ago), licensed Apache-2.0. It adds 143 tokens to every session and 1,695 once invoked, about $0.0007 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.