ennodia 0.1.1 → 0.2.0

package/docs/guides/running-better-audits.md

@@ -1,24 +1,26 @@
 ---
-title: Running Better Audits
-description: How to ask Ennodia for audits that catch user-facing product problems, not only consistency issues.
+title: Better Audits
+description: How Ennodia audits can detect product problems that users see, not only consistency issues.
 ---
 
-Ennodia can ask several agents for review, but the result still depends on the
-audit rubric. A vague request such as "audit the website" often turns into a
-consistency pass: broken links, mismatched docs, release metadata, type errors,
-and copy drift. That is useful, but it does not prove the page is clear to a
-first-time visitor.
+Ennodia can request review from multiple agents. The result still depends on the
+audit rubric. A vague request such as "audit the website" often produces a
+consistency pass.
+
+This pass can find broken links, mismatched documentation, release metadata errors, type
+errors, and copy drift. It does not prove that the page is clear to a first-time
+visitor.
 
 ## Quick Recipe
 
-Use a source-grounded audit when the answer depends on docs, standards, package
-behavior, or another product's public surface:
+Use a source-grounded audit for answers that depend on documentation,
+standards, package behavior, or another product's public surface:
 
 ```json
 {
   "tool": "ennodia_run",
   "arguments": {
-    "prompt": "Audit these docs against the linked primary sources. Separate facts from judgment and recommend exact edits.",
+    "prompt": "Audit this documentation against the linked primary sources. Separate facts from judgment and recommend exact edits.",
     "mode": "parallel",
     "compare": true,
     "skillIds": ["source-grounded-audit"]
@@ -26,8 +28,8 @@ behavior, or another product's public surface:
 }
 ```
 
-Install the skill first if the harness has not seen it yet. See
-[Using Agent Skills](/docs/guides/agent-skills/).
+Install the skill first if the harness does not have it. See
+[Agent Skills](/docs/guides/agent-skills/).
 
 ## Pick the Audit Mode
 
@@ -35,63 +37,67 @@ Name the failure mode you want reviewers to catch.
 
 | Audit mode | Main question |
 | --- | --- |
-| Consistency | Do the docs, website, package metadata, and behavior agree? |
+| Consistency | Do the documentation, website, package metadata, and behavior agree? |
 | Conversion | Would a first-time visitor understand why to install this? |
-| Visual QA | Does the rendered page look balanced on desktop and mobile? |
+| Visual quality assurance (QA) | Does the rendered page look balanced on desktop and mobile? |
 | Accessibility | Can the page be navigated and understood with assistive technology? |
 | Release | Can the package be published, installed, and launched safely? |
 
 Use separate passes when the stakes are different. A page can pass consistency
 and still fail conversion.
 
-For a large review, split the work into focused slices instead of sending the
-entire prompt to every reviewer. See
+For a large review, split the work into focused slices. Do not send the entire
+prompt to each reviewer. See
 [Compositional Audits](/docs/concepts/compositional-audits/).
 
 ## Landing Page Rubric
 
-For landing pages, ask each reviewer to answer these questions:
+For landing pages, require each reviewer to answer these questions:
 
 - Can a first-time visitor explain what the product does within ten seconds?
 - Is the reason to install visible in the hero or first section?
 - Is the install path visible before architecture details?
-- Does the page use user language instead of maintainer language?
+- Does the page use user language and avoid maintainer language?
 - Are screenshots or rendered viewports available for visual judgment?
-- Does the final section strengthen trust instead of ending on defensive copy?
+- Does the final section strengthen trust and avoid defensive copy?
 
-For Ennodia itself, prefer `mode: "parallel"` and `compare: true` for this kind
-of review. A single source-only answer is not enough for product-facing pages.
+For Ennodia, use `mode: "parallel"` and `compare: true` for this type of review.
+A single source-only answer is not sufficient for product pages.
 
 ## Include Rendered Evidence
 
-Source-only review misses visual problems. Before asking Ennodia to judge a
-website, include desktop and mobile screenshots or ask a harness with browser
+Source-only review misses visual problems. Before an Ennodia website review,
+include desktop and mobile screenshots. You can also use a harness with browser
 access to inspect the rendered page.
 
-Rendered evidence catches problems such as oversized logos, unbalanced hero
-spacing, text that feels too large or too small, delayed content, and accidental
-visual hierarchy.
+Rendered evidence can show oversized logos, unbalanced hero space, incorrect
+text size, delayed content, and accidental visual hierarchy.
+
+When local files matter, include the file paths and a short text description.
+Then inspect the child outputs for access errors.
 
-When screenshots or other local files matter, include both the file paths and a
-short text description. Then check the child outputs for access errors. If a
-harness cannot read the files, stage them somewhere it can access and rerun; do
-not treat that failure as a normal design review.
+If a harness cannot read the files, put them in an accessible location and run
+the task again. Do not treat that failure as a normal design review.
 
 ## Ground Standards in Sources
 
-When the audit depends on an external standard, registry convention, or product
-behavior, include the sources in the prompt and ask reviewers to separate facts
-from judgment. Otherwise a model council can confidently optimize the wrong
-model of the world.
+Include primary sources when an audit depends on an external standard, registry
+convention, or product behavior. Require reviewers to separate facts from
+judgment.
+
+Without these sources, a model council can confidently use an incorrect model
+of the world.
+
+Agent Skills are one cautionary example. They are not an Ennodia-private prompt
+format. They are portable `SKILL.md` folders.
 
-Agent Skills are the cautionary example: they are not an Ennodia-private prompt
-format. They are portable `SKILL.md` folders discovered by tools such as Codex,
-Claude Code, OpenCode, and Antigravity in native locations. A useful audit for
-skills must start from those product docs before recommending an Ennodia API.
+Tools such as Codex, Claude Code, OpenCode, and Antigravity discover these
+folders in native locations. Reviewers must read that product documentation
+before they recommend an Ennodia application programming interface (API).
 
-The bundled `source-grounded-audit` skill exists for this case. Install it into
-the harnesses you use for standards-sensitive audits, then ask Ennodia to use
-that skill when the answer depends on external docs.
+The bundled `source-grounded-audit` skill supports this case. Install it in the
+harnesses that perform standards-sensitive audits. Then tell Ennodia to use the
+skill when the answer depends on external documentation.
 
 ## Good Prompt Shape
 
@@ -107,11 +113,11 @@ If screenshots are available, judge the rendered page too.
 
 Return:
 1. The top conversion blockers, prioritized.
-2. Whether the first viewport explains what this is and why to install it.
+2. If the first viewport explains what this is and why to install it.
 3. Exact replacement copy for weak sections.
 4. Visual changes needed for desktop and mobile.
 5. Which issues a consistency-only audit would miss.
 ```
 
-The last item matters. It forces the reviewer to state the limits of the audit,
-which helps decide whether another pass is needed.
+The last item makes the reviewer state the audit limits. This information helps
+you decide if another pass is necessary.

package/docs/index.md

@@ -1,72 +1,89 @@
 ---
 title: What is Ennodia?
-description: Start here to understand Ennodia, choose the right setup path, and find the docs page for your task.
+description: Start here to understand Ennodia, select an installation path, and find the documentation page for your task.
 ---
 
-Ennodia is a local MCP server for multi-agent review. Your primary AI agent can
-ask other installed agent CLIs for help, inspect what happened, and let a model
-compare the answers before you get one result.
+Ennodia is a local Model Context Protocol (MCP) server for multi-agent review.
+Your primary artificial intelligence (AI) agent can request help from installed
+agent command-line interfaces (CLIs).
 
-It is useful when one agent is already working and the task is important enough
-to spend extra model work on review, planning, diagnosis, or comparison.
-Ennodia is deliberation-class tooling: a run usually takes minutes, and Compare
-adds a serial judge pass plus a serial synthesizer pass after child agents
-finish.
+The primary agent can inspect the work. It can use a Judge and Result Advisor to
+compare the answers.
+
+An optional Plan Advisor can first propose a bounded team. It cannot execute the
+proposal.
+
+Use Ennodia when one agent has a task that needs additional review, a plan, a
+diagnosis, or a comparison. A run usually takes minutes.
+
+Compare adds a serial Judge pass and a serial Result Advisor pass after the
+child agents finish.
 
 ## The Problem Ennodia Solves
 
-The primary problem is single-model judgment as a single point of failure. One
-agent can miss the risky bug, accept a weak plan, or hallucinate confidence.
-For important work, you often want independent second opinions and a record of
-what each reviewer actually did.
+Single-model judgment can be a single point of failure. One agent can miss a
+risky bug, accept a weak plan, or report false confidence.
+
+Important work can require independent second opinions and a record of each
+reviewer's work.
 
-The enabling asset is that many developers already pay for several independent
-agent subscriptions — Codex, Claude Code, Antigravity, OpenCode, and others.
-Ennodia turns the installed CLIs on one machine into a review panel your
-primary agent can use, without new accounts, API keys, or vendor migration:
+Many developers have independent agent subscriptions. Examples include Codex,
+Claude Code, Antigravity, and OpenCode. Ennodia makes their installed CLIs a
+review panel for the primary agent.
 
-- Agents reach Ennodia through MCP, the protocol agents already speak.
-- Model-led Compare produces a structured disagreement map and final synthesis.
+This process does not require new accounts or a vendor migration. It does not
+require new application programming interface (API) keys.
+
+- Agents reach Ennodia through MCP, a protocol that the agents support.
+- A Plan Advisor can propose inert harness, model, and skill assignments. A
+  separate explicit call must launch the validated plan.
+- Model-led Compare uses a Judge for the disagreement map and a Result Advisor
+  for the recommended answer.
 - One shared core discovers harnesses, plans routes, checks local preflight
   limits, tracks tasks, and keeps the trace visible.
-- The separate Ennodia IO package is an experimental app-facing appendix.
+- The separate Ennodia IO package is an experimental appendix for apps.
 
 ## Why People Use It
 
 Ennodia:
 
 - discovers supported local harnesses automatically
-- uses the CLIs, installs, subscriptions, and model choices you already have
+- uses your installed CLIs, subscriptions, and model choices
 - estimates preflight input-token budgets and enforces local caps on that
   estimate
-- keeps child task IDs, status, ETA, output, failures, Compare state, and run
-  history visible, including terminal receipts across server restarts
-- lets a model judge and synthesize parallel answers
+- keeps child task identifiers (IDs), status, estimated completion time,
+  output, failures, Compare state, and run
+  history visible
+- can retain terminal receipts across server restarts
+- lets a Judge inspect parallel answers and a Result Advisor combine the
+  findings
 - uses native Agent Skills through harness-visible `SKILL.md` folders
 
-Durable model and skill preference memory is roadmap work. Terminal run history
-is persisted locally under `~/.ennodia/history/` by default; live in-progress
-state remains process-local.
+Durable model and skill preference memory is roadmap work. By default, Ennodia
+stores terminal run history under `~/.ennodia/history/`. Live in-progress state
+remains process-local.
 
 ## Choose Your Path
 
-| If you want to... | Start here |
+| Task | Page |
 | --- | --- |
-| Ask an agent to install Ennodia for you | [Installation for Agents](./install/) |
-| Set it up manually or work from a checkout | [Quickstart](./getting-started/) |
+| Request Ennodia installation from an agent | [Installation for Agents](./install/) |
+| Install it manually or work from a checkout | [Quickstart](./getting-started/) |
 | Estimate or limit a costly run | [Budgets and Limits](./guides/budgets-and-limits/) |
-| Install or use `source-grounded-audit` and other skills | [Using Agent Skills](./guides/agent-skills/) |
+| Install or use `source-grounded-audit` and other skills | [Agent Skills](./guides/agent-skills/) |
 | Understand the run lifecycle | [How Ennodia Works](./concepts/how-ennodia-works/) |
+| Request a bounded team proposal before work | [How Ennodia Works](./concepts/how-ennodia-works/#plan-advisor) |
 | Use panels for second opinions | [Second Opinions as Infrastructure](./concepts/second-opinions/) |
 | Understand local storage and data movement | [Data Governance](./concepts/data-governance/) |
 | Understand Core, <span class="agent-logo agent-logo--mcp" aria-hidden="true"></span>MCP, and experimental IO | [Interfaces and Core](./concepts/interfaces-and-core/) |
 | Split a large review into smaller shards | [Compositional Audits](./concepts/compositional-audits/) |
 | Call the exact <span class="agent-logo agent-logo--mcp" aria-hidden="true"></span>MCP tools | [MCP Tools](./reference/mcp-tools/) |
-| Call Ennodia from a local HTTP app | [Ennodia IO](./reference/ennodia-io/) |
-| Check harness IDs and setup notes | [Supported Harnesses](./reference/supported-harnesses/) |
+| Call Ennodia from a local web app | [Ennodia IO](./reference/ennodia-io/) |
+| Check harness identifiers and installation notes | [Supported Harnesses](./reference/supported-harnesses/) |
+| Use the ASD-STE100 rules and approved Ennodia terms | [Controlled English](./reference/controlled-english/) |
 | Inspect benchmark receipts | [Benchmarks](./reference/benchmarks/) |
-| Decide whether Ennodia fits your workflow | [Comparisons](./comparisons/) |
-| Get better output from model-led Compare | [Running Better Audits](./guides/running-better-audits/) |
+| Decide if Ennodia fits your workflow | [Comparisons](./comparisons/) |
+| Get better output from model-led Compare | [Better Audits](./guides/running-better-audits/) |
 
 ## First Useful Run
 
@@ -77,5 +94,11 @@ state remains process-local.
 5. Poll `ennodia_get_run` until it reaches `succeeded`, `failed`, or
    `cancelled`.
 
-The user should not need to manually grade every child answer unless they want
-to inspect the trace.
+The user does not normally need to grade each child answer. The trace remains
+available for inspection.
+
+For a team proposal, call `ennodia_start_plan_advice`. Then poll
+`ennodia_get_plan_advice`. A ready result is only validated data.
+
+Call `ennodia_start_advised_plan` with the returned plan digest once to start the
+work.

package/docs/install.md

@@ -1,11 +1,10 @@
 ---
 title: Installation for Agents
-description: A copy-paste setup page for agents installing Ennodia into Codex, Claude Code, OpenCode, Antigravity, and other MCP clients.
+description: A copy-paste installation page for agents that install Ennodia in Codex, Claude Code, OpenCode, Antigravity, and other Model Context Protocol (MCP) clients.
 ---
 
-This page is for the agent doing the setup. If the user gave you this link, the
-goal is to install Ennodia as a local MCP server and turn it on in the MCP
-client they use.
+This page is for the agent that does the installation. Install Ennodia as a local Model
+Context Protocol (MCP) server. Then enable it in the user's MCP client.
 
 ## Copy-Paste Prompt
 
@@ -15,18 +14,18 @@ try-ennodia.cherninlab.com
 
 ## Install Command
 
-Use the npm package for normal MCP client setup:
+Use the npm package for normal MCP client installation:
 
 ```sh
 npx -y ennodia
 ```
 
-Requires Bun `1.3.14` or newer. `npx` downloads Ennodia; Bun runs it. If the
-client supports Bun directly, `bunx ennodia` is also fine.
+Ennodia requires Bun `1.3.14` or newer. `npx` downloads Ennodia. Bun runs it.
+If the client supports Bun directly, use `bunx ennodia`.
 
-## MCP Config
+## MCP Configuration
 
-Add this server to the user's MCP client config:
+Add this server to the user's MCP client configuration:
 
 ```json
 {
@@ -54,7 +53,7 @@ For a local Ennodia checkout, use:
 
 ## First Agent Checks
 
-After configuring MCP, call these tools:
+After MCP configuration, call these tools:
 
 1. `ennodia_list_harnesses`
 2. `ennodia_estimate_budget`
@@ -62,35 +61,44 @@ After configuring MCP, call these tools:
 4. `ennodia_run`
 5. `ennodia_get_run`
 
-`ennodia_list_harnesses` should show which local tools are available — see the
-current adapter list and IDs in
+`ennodia_list_harnesses` shows the available local tools. See the current
+adapter list and identifiers (IDs) in
 [Supported Harnesses](/docs/reference/supported-harnesses/).
 
-## What You Are Installing
+## Installed Components
 
-Ennodia is a local MCP server. It lets the user's main agent ask other installed
-agent CLIs for help during the same task.
+Ennodia is a local MCP server. It lets the user's main agent request help from
+other installed agent command-line interface (CLI) programs during the same task.
 
-The important part is not "many agents" by itself. Ennodia works with existing
-local installs, subscriptions, and model choices. It discovers supported CLIs,
-starts child tasks, tracks status and output, estimates the input-token budget,
-and can ask a model to compare the answers before returning one result.
+The number of agents is not the primary benefit. Ennodia uses current local
+installs, subscriptions, and model choices. It discovers supported CLI programs and
+starts child tasks. It also tracks status and output.
 
-Ennodia is asynchronous by design. A primary agent can start a run, keep working,
-and later poll `ennodia_get_run` for exact status, ETA, child task IDs, Compare
-state, failures, and the final answer.
+Ennodia estimates the input-token budget. A Judge can compare the answers. A
+Result Advisor can then recommend one result.
+
+An optional Plan Advisor can propose harness, model, and skill assignments. It
+cannot start them. A separate explicit call must start the validated plan.
+
+Ennodia is asynchronous by design. A primary agent can start a run and continue
+its main task. It can later poll `ennodia_get_run` for status and estimated
+completion time.
+
+The result also contains child task IDs, Compare state, failures, and the final
+answer.
 
 Tell the user a real run can take minutes. Compare adds two serial model passes
 after the selected child agents finish.
 
 ## Example Requests
 
-Use Ennodia when the user is intentionally spending more model work to get a
-better answer than one agent usually gives.
+Use Ennodia when the user permits more model work to get a better answer than
+one agent usually gives.
 
 ```text
-Use Ennodia to ask several available agents to review this bug fix. Compare the
-answers, name disagreements, and return the highest-risk issue first.
+Use Ennodia to request a review from multiple available agents.
+Compare the answers for this bug fix.
+Name the disagreements and return the highest-risk issue first.
 ```
 
 ```text
@@ -99,19 +107,20 @@ Compare the plans and tell me which assumptions each model made.
 ```
 
 ```text
-Use Ennodia twice: once with source-grounded-audit and once without it. Compare
-what the skill changed before recommending the final docs edit.
+Use Ennodia twice.
+Use source-grounded-audit for the first run, but not for the second run.
+Compare the results before you recommend the final documentation edit.
 ```
 
-Compare is model-led. The user should not have to manually judge parallel
-answers unless they want to inspect the trace.
+Compare is model-led. The user does not normally need to judge the parallel
+answers. The trace remains available for inspection.
 
 ## Optional Next Steps
 
 - Use [Budgets and Limits](/docs/guides/budgets-and-limits/) before costly
   parallel runs.
-- Use [Using Agent Skills](/docs/guides/agent-skills/) before passing
+- Use [Agent Skills](/docs/guides/agent-skills/) before you pass
   `skillIds` such as `source-grounded-audit`.
 - Use [Supported Harnesses](/docs/reference/supported-harnesses/) when a local
-  CLI is missing, unrunnable, or using the wrong model ID.
+  CLI is missing, not runnable, or has the wrong model ID.
 - Use [MCP Tools](/docs/reference/mcp-tools/) for exact parameter shapes.

package/docs/reference/benchmarks.md

@@ -4,8 +4,8 @@ description: Deterministic benchmark snapshots for Ennodia's multi-agent bug-rec
 ---
 
 The first benchmark is `multi-model-bug-recall`: TypeScript review fixtures with
-committed bug oracles. It measures whether a condition recalls required
-findings and avoids known false-positive traps.
+committed bug oracles. It measures required-finding recall for a condition and
+checks for known false-positive traps.
 
 ## Current Snapshot
 
@@ -19,6 +19,8 @@ bun run bench:bug-recall -- --out /tmp/ennodia-benchmark-snapshot
 
 Mode: fixture, using committed reference outputs.
 
+The F1 score combines recall and precision. FP means false positive.
+
 | Condition | Cases | Recall | Precision | F1 | High recall | FP traps |
 | --- | ---: | ---: | ---: | ---: | ---: | ---: |
 | `claude-code-solo` | 4 | 87.5% | 100% | 93.3% | 100% | 0 |
@@ -47,8 +49,8 @@ does not launch child agents:
 bun run bench:bug-recall
 ```
 
-Live mode starts real local harnesses and varies with installed CLI versions,
-models, account state, and machine configuration:
+Live mode starts real local harnesses and varies with installed command-line
+interface (CLI) versions, models, account state, and machine configuration:
 
 ```sh
 bun run bench:bug-recall:live -- --fixture 001-missing-await

package/docs/reference/controlled-english.md

@@ -0,0 +1,130 @@
+---
+title: Controlled English
+description: Ennodia terminology and writing rules for ASD-STE100 Issue 9.
+---
+
+# Controlled English
+
+Ennodia documentation uses ASD-STE100 Issue 9 as its writing standard.
+This policy applies to the README, documentation, website text, and visible asset text.
+
+Commands, API fields, status values, model IDs, and quoted interface text are data.
+Keep this data exact, even when its text does not use controlled English.
+
+## Writing Rules
+
+- Use American English.
+- Use active voice when the agent is known.
+- Give one instruction in each procedural sentence.
+- Use 20 words or fewer in a procedural sentence.
+- Use 25 words or fewer in a descriptive sentence.
+- Use six sentences or fewer in one paragraph.
+- Do not use semicolons.
+- Do not use a contraction.
+- Use one approved term for each concept.
+- Define a necessary abbreviation at its first use on each independent page.
+- Preserve exact product names, commands, paths, model IDs, and API fields.
+
+The landing page uses `AI makes mistakes` as an approved fixed tagline.
+Define `artificial intelligence (AI)` in the sentence directly below it.
+
+Automated checks can find some defects, but they cannot certify compliance.
+A human reviewer must confirm meaning, grammar, and correct technical use.
+
+## Ennodia Termbase
+
+The Ennodia project approves the following terms for its software documentation.
+Use the listed meaning and part of speech.
+
+| Term | Type | Approved meaning |
+| --- | --- | --- |
+| Ennodia | technical noun | The software product in this repository. |
+| artificial intelligence (AI) | technical noun | Software that produces or evaluates model output. |
+| AI makes mistakes | approved tagline | The exact Ennodia landing-page headline. Keep this phrase unchanged. |
+| Model Context Protocol (MCP) | technical noun | The protocol that exposes Ennodia tools to a compatible client. |
+| command-line interface (CLI) | technical noun | A text interface that accepts commands. |
+| application programming interface (API) | technical noun | A defined interface that software uses to exchange data. |
+| Hypertext Transfer Protocol (HTTP) | technical noun | The protocol that Ennodia IO uses for local requests. |
+| user interface (UI) | technical noun | The visible controls and output that a user can operate. |
+| Ennodia IO | product name | The separate local HTTP package for Ennodia. |
+| harness | technical noun | An Ennodia adapter that starts one supported local CLI. |
+| model | technical noun | A named AI model that a harness can select. |
+| large language model (LLM) | technical noun | A model that processes and produces language or code. |
+| Agent Skill | technical noun | Installed instructions that a compatible harness can load. |
+| skill | technical noun | The short form of Agent Skill after the full term appears. |
+| task | technical noun | One harness process that Ennodia starts and tracks. |
+| worker | technical noun | A task that performs one assigned part of a plan. |
+| worker pool | technical noun | A defined group of workers with the same assignment type. |
+| slice | technical noun | One focused part of a compositional task. |
+| plan | technical noun | Validated data that defines ordered worker tasks. |
+| plan digest | technical noun | A hash that identifies the exact validated plan. |
+| inventory | technical noun | The runnable harnesses and installed skills that Ennodia finds. |
+| inventory fingerprint | technical noun | A hash that identifies one inventory state. |
+| Plan Advisor | technical noun | The role that proposes a bounded plan as inert data. |
+| Judge | technical noun | The Compare role that maps agreements, conflicts, and risks. |
+| Result Advisor | technical noun | The Compare role that recommends an answer from available evidence. |
+| Compare | technical noun or verb | The Ennodia operation that uses a Judge and a Result Advisor. |
+| run | technical noun or verb | A tracked orchestration, or the action that starts it. |
+| receipt | technical noun | A bounded record of tasks, events, results, and execution data. |
+| route | technical noun or verb | A harness selection, or the action that selects a harness. |
+| poll | technical verb | To request current status at controlled intervals. |
+| review | technical noun or verb | An inspection of code, documentation, or model output for defects. |
+| standard input/output (stdio) | technical noun | The paired input and output streams of an MCP server process. |
+| standard input (stdin) | technical noun | The input stream of a local process. |
+| standard output (stdout) | technical noun | The normal output stream of a local process. |
+| standard error (stderr) | technical noun | The error output stream of a local process. |
+| token | technical noun | A model input or output unit. |
+| budget | technical noun | A caller limit for tasks or estimated input tokens. |
+| provider | technical noun | A company or service that supplies a model or harness. |
+| subscription | technical noun | A provider plan that can set access or usage limits. |
+| quota | technical noun | A provider limit on available usage. |
+| estimated time of arrival (ETA) | technical noun | An estimate of the time until a task ends. |
+| Cross-Origin Resource Sharing (CORS) | technical noun | An HTTP mechanism that controls browser access from another origin. |
+| bring your own key (BYOK) | technical noun | An app pattern that uses credentials supplied by the user. |
+| Codex | product name | The OpenAI coding-agent product. |
+| Claude Code | product name | The Anthropic coding-agent product. |
+| Antigravity | product name | The Google coding-agent CLI used by Ennodia. |
+| Gemini | product name | A Google model family. |
+| GPT | product name | An OpenAI model family. |
+| OpenCode | product name | A supported coding-agent CLI. |
+| Kilo Code | product name | A supported coding-agent CLI. |
+| Hermes Agent | product name | A supported coding-agent CLI. |
+| Kiro CLI | product name | A supported coding-agent CLI. |
+| Cline | product name | A supported coding-agent CLI. |
+| AutoGen | product name | An external multi-agent framework. |
+| ChatHub | product name | An external multi-model comparison product. |
+| LangGraph | product name | An external agent workflow framework. |
+| OpenRouter | product name | An external model routing service. |
+| Mixture-of-Agents | technical noun | An external multi-model inference method. |
+| model merging | technical noun | A method that combines model parameters. |
+| fine-tuning | technical noun | A method that adapts a model with additional training. |
+| mergekit | product name | An external toolkit for model merging. |
+| npm | product name | The package registry and command-line tool for JavaScript packages. |
+| JavaScript Registry (JSR) | product name | The registry that distributes the Ennodia JavaScript package. |
+| Model Context Protocol Registry | product name | The official registry for MCP server metadata. |
+| GitHub Actions | product name | The service that runs Ennodia release and deployment workflows. |
+| Cloudflare Pages | product name | The service that hosts the Ennodia website. |
+| Astro | product name | The framework that builds the Ennodia website. |
+| Starlight | product name | The Astro documentation framework used by the Ennodia website. |
+| Semantic Versioning (SemVer) | technical noun | The version format that identifies compatible and incompatible releases. |
+| JSON Lines | technical noun | A text format that stores one JSON value on each line. |
+| JavaScript Object Notation (JSON) | technical noun | A structured text format for data exchange. |
+| YAML | technical noun | A structured text format that Ennodia uses for Agent Skill metadata. |
+| OpenID Connect (OIDC) | technical noun | The identity protocol used for tokenless registry publication from GitHub Actions. |
+| identifier (ID) | technical noun | A value that uniquely identifies an Ennodia object. |
+| F1 score | technical noun | The harmonic mean of precision and recall in a benchmark. |
+| false positive (FP) | technical noun | A reported finding that the benchmark oracle does not require. |
+| TypeScript | product name | The typed programming language used by Ennodia. |
+| benchmark | technical noun or verb | A repeatable quality test, or the action that runs that test. |
+
+Use `Plan Advisor` and `Result Advisor` in full.
+Do not use `Advisor` alone when the role can be unclear.
+
+Legacy fields can contain `synthesizer` or `synthesis` in their exact names.
+Use `Result Advisor` for the current public role.
+
+## Source
+
+Use the official [ASD-STE100 Issue 9 standard](https://www.asd-ste100.org/assets/files/ASD-STE100_ISSUE9.pdf).
+The ASD Simplified Technical English Maintenance Group controls that standard.
+Do not copy its dictionary into this repository.