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.
git clone --depth 1 https://github.com/maxime-tremblay/APEXLang-Project-TemplateWrote 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/agents/maxime-tremblay/apexlang-project-template/40-agent-revision)<a href="https://agentmods.dev/agents/maxime-tremblay/apexlang-project-template/40-agent-revision"><img src="https://agentmods.dev/badge/agents/maxime-tremblay/apexlang-project-template/40-agent-revision/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/agents/maxime-tremblay/apexlang-project-template/40-agent-revision"><img src="https://agentmods.dev/badge/agents/maxime-tremblay/apexlang-project-template/40-agent-revision.svg" alt="Reviewed on agentmods" width="80" 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 | $0.00000 | $0.06022 |
| Opus 5 | $0.00000 | $0.03011 |
| Sonnet 5 | $0.00000 | $0.01204 |
| Haiku 4.5 | $0.00000 | $0.00602 |
Grade A, and why
40-agent-revision 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 10d 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 — 232 lines — stays where its author put it; the contents beside it link to each section on GitHub.
All
node tools/apexctl.mjs ...commands are package-root relative: run them from the packaged skill root, or invoke that script by explicit path.
Agent 3 — Revision (Child Workflow)
Purpose
- Apply accepted review notes to produce the final APEXlang artifacts, enforce governance gates, and persist only compact runtime evidence when needed — with minimal token/context usage.
Inputs
- draft_path: path to the app file in the transient temp workspace outside the repo
- critique_path: internal review payload reference for the current run
- target_type
- output_path? (defaults per scope)
- app_root? (when Whole Application scope is active; e.g., applications/app_###/)
- server_side_condition? (structured object from master workflow)
<authority_rules>
- Shared prompt contract:
references/workflows/apexlang/prompt-contracts.md - Preserve its hierarchy, rule IDs, intermediate artifacts, and stop conditions while applying accepted fixes. </authority_rules>
<task_scope>
- Apply accepted critique notes only.
- Keep revision deterministic and avoid re-planning unless the critique explicitly requires a plan repair. </task_scope>
<generation_plan_contract>
- Preserve the frozen
Generation Planwhen applying accepted notes. - If an accepted note changes a plan-level structural decision, record a
Generation Plan Repair, update the relevant plan fields first, then reconcile the emitted artifact to that repaired plan. - Do not finalize a revision that still violates
GENERATION_PLAN_DRIFT_001. </generation_plan_contract>
<output_contract>
- Preserve or refresh
compiler_truth_evidencewhen structural revisions change a compiler-justified decision. - Keep revision output aligned with the same response-order contract used by Draft when an updated plan or evidence block must be emitted. </output_contract>
<stop_conditions>
- Follow
RULES_FIRST_WORKFLOW_REQUIRED_001andHUMAN_INTERVENTION_REQUIRED_001. - If deterministic revision cannot satisfy the shared contract without inventing structure, keep finals blocked and record Missing Inputs instead of guessing. </stop_conditions>
Responsibilities (apply only accepted notes)
- Respect governance:
- Reject any change conflicting with references/policies/memory-bank/00-guard/ai.guard.md or 10-global/apex.global.md.
- Do not invent attributes or CSS classes; use templates/* only.
- Reject finals that attempt to copy content or reusable patterns from
applications/**; regenerate using the canonical template instead. - Preserve target-app reads only when they provide concrete integration facts such as existing ids, aliases, navigation entries, breadcrumb entries, or artifact paths.
- Do not create unrelated helper source files for APEXlang generation tasks unless the user explicitly requested tooling or scripts.
- For dynamic-action revisions, follow
references/domains/business-logic/dynamic-actions/workflow-dynamic-actions.mdto maintain naming, template selection, and itemsToSubmit discipline. For batch runs, iterate through canonicaltargets(legacytarget_pagesallowed) and ensure each page’s identifiers (regions/items/buttons) are updated before writing finals. - For batch revisions across all domains, normalize legacy inputs (
target_pages,target_items,target_buttons,target_button,apply_to) to canonicaltargets, then apply deterministic per-target outcomes. - For mixed translation/control prompts, do not finalize localization artifacts until runtime language switching versus text-message localization intent has been resolved.
- Page processes: use
references/domains/business-logic/processes/workflow-page-processes-batch.mdfor invokeApi batches, updating each page’s button guards and parameter blocks accordingly. - Application processes: enforce
type: executeCodeonly and removeinvokeApiusage fromappProcessblocks. - Item computations: leverage
references/domains/business-logic/computations/workflow-computations-batch.mdto apply template outputs per target item, retain only compact run evidence when needed, and convert logic to packaged APIs where recommended.- For translations, keep placeholders intact and use
references/domains/shared-components/workflow-translations.md; for bundles, iterate viareferences/domains/shared-components/workflow-translations-batch.md, logging AI-assisted entries separately.
- For translations, keep placeholders intact and use
- For button batches, apply
references/domains/page-components/buttons/workflow-button-batch.md, ensuring layout slots, action targets, nested trigger actions, and confirmations align with guardrails. Persist only compact run evidence when needed. - For non-button batch domains (processes, computations, DA batches, translations, SSC/help-text), ensure compact run evidence includes
operation, canonicaltargets, and per-target statuses before writing finals when the workflow explicitly persists logs. - Deterministic button autofix: when critique reports a button-behavior policy rule violation, apply the JSON policy fix from
assets/component-policies.jsonexactly. ForBTN_RULE_001, removewarnOnUnsavedChangeswheneverbehavior.action: definedByDynamicAction. - Deterministic button templateOptions autofix: when critique or runtime reports button template-option alias drift, preserve
#DEFAULT#, keep multi-value arrays bracketed with one accepted value per line, and normalize aliases/naked suffixes to canonical emitted values from the button family contract. Normalizeleft|iconLeft|icon-left -> t-Button--iconLeft,right|iconRight|icon-right -> t-Button--iconRight,push|hoverIconPush|hover-icon-push -> t-Button--hoverIconPush,spin|hoverIconSpin|hover-icon-spin -> t-Button--hoverIconSpin,hide-icon-on-desktop|desktopHideIcon -> t-Button--desktopHideIcon,hide-label-on-mobile|mobileHideLabel -> t-Button--mobileHideLabel,primary -> t-Button--primary,simple -> t-Button--simple,tiny -> t-Button--tiny,stretch -> t-Button--stretch,pillStart|pill-start -> t-Button--pillStart,padLeft|pad-left -> t-Button--padLeft,gapRight|gap-right -> t-Button--gapRight,padTop|pad-top -> t-Button--padTop,gapBottom|gap-bottom -> t-Button--gapBottom,link -> t-Button--link,success -> t-Button--success, andnoUI|no-ui -> t-Button--noUI. - Deterministic display-only autofix:
- For
pageItemwithtype: displayOnly, remove unsupportedappearance.widthandappearance.height. - Ensure
settings.sendOnPageSubmit: falseexists; insertsettings { sendOnPageSubmit: false }when missing or normalize value when not false. - Keep all other display-only attributes unchanged unless critique explicitly requests additional revisions.
- For
- Deterministic master-detail Content Row autofix:
- For
CONTENT_ROW_SETTINGS_SUBSTITUTION_REQUIRED_001, rewrite Content Rowsettings.overline,settings.title,settings.description, andsettings.miscellaneouscolumn mappings from bare aliases to&COLUMN_NAME.substitutions, but leave literal values such asoverline: Employeeunchanged. - For
CONTENT_ROW_SELECTION_ITEMS_REQUIRED_001, add or fix native Content Row selection items. ForfocusOnly, emit onlyrowSelection { type: focusOnly }and remove selection page-item references. ForsingleSelection, emitrowSelection { type: singleSelection currentSelectionPageItem: Pn_SELECTED_KEY }and create the same-page hidden item. FormultipleSelection, emitrowSelection { type: multipleSelection currentSelectionPageItem: Pn_SELECTED_KEYS selectAllPageItem: Pn_SELECT_ALL }, create the same-page hidden current-selection item, and create the same-page select-all checkbox item. - For
MASTER_DETAIL_CONTENT_ROW_ACTION_REQUIRED_001, add or fix a report-mode Content Rowaction ... (withposition: fullRowLinkand a structured same-page target that sets the hidden context item from the parent PK using&COLUMN.substitution. Do not accept nativerowSelection.currentSelectionPageItemas sufficient for master-detail context setting. - For
MASTER_DETAIL_VISIBLE_SELECTOR_REGRESSION_001, convert the visible parent select-list selector into a protected hidden context item unless the user explicitly requested manual selection. - For
MASTER_DETAIL_CHILD_BIND_SUBMIT_REQUIRED_001orIR_CONTEXT_BIND_SUBMIT_REQUIRED_001, add every same-page item referenced by child Interactive Report SQL binds tosource.pageItemsToSubmit. - For
MASTER_DETAIL_TOOLBAR_ACTIONS_REQUIRED_001, move child create/edit/detail buttons into the child Interactive Report scope usinglayout.region: @<child-report-static-id>andslot: rightOfInteractiveReportSearchBar. - For
MASTER_DETAIL_LAYOUT_REQUIRED_001, use the master-detail Content Row layout recipe: parent Content Row first withcolumnSpan: 3or4, child report second withstartNewRow: false, and remove redundant childcolumn/columnSpanunless the validated contract requires them. - For
IR_PROJECTED_COLUMNS_REQUIRED_001, add missing Interactive Reportcolumn (...)definitions for every SQL projection usinginteractive-report._columns._common.md; preserve numeric end alignment and required headings.
- For
- Deterministic layout autofix:
- Apply these fixes to generated application artifacts before final validation; app output is not a layout-lint bypass.
- Apply layout fixes per local scope: page slot rows, nested
parentRegionrows, item rows bylayout.region + layout.slot, and button rows bylayout.region + layout.slot. - For equal-width sibling rows, remove
layout.columnandlayout.columnSpan. - Ensure the first component in an equal-width row omits
layout.startNewRow. - Ensure second-and-later equal-width siblings set
layout.startNewRow: false. - Do not alter intentionally asymmetric rows such as sidebar-main, faceted-search, or parent-child split layouts.
- Finals:
- Write revised
.apxoutputs only to the transient temp app:- Pages → temp
{output_path}/pages/or{app_root}/pages/ - Shared components → temp
{output_path}/shared-components/or{app_root}/shared-components/
- Pages → temp
- Publish into
applications/<target-app>/...only after the resolved live runtime action succeeds.
- Compact run evidence:
- Persist a concise list of applied/rejected/deferred notes only when the workflow explicitly stores durable evidence under
the temp-runtime logs directory underAPEXLANG_OUTPUT_ROOT/logs/``. - Preserve
compiler-truth-report.jsonfor generated or revised.apxartifacts and block completion when the report is missing or failing. - Preserve
validation-report.json,validation-transcript.log,problems.json, andcomponent-contracts/<build>.jsonfromruntime validate. Apply fixes only for reported validation problems, then rerunruntime validateuntil live/compiler evidence and VSCode Problems evidence pass. - Resolve or explicitly defer parallel-skill contradictions recorded by critique; unresolved conflicts must remain as
Missing Inputs.
- Persist a concise list of applied/rejected/deferred notes only when the workflow explicitly stores durable evidence under
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.
- 10d ago First seen · 232 lines · 0 tokens per session scan A 334614d463fa
40-agent-revision is an agent published in the GitHub repository maxime-tremblay/APEXLang-Project-Template (2 stars, last pushed 3mo ago), licensed MIT. It costs nothing until one of its globs matches a file; then it loads 6,022 tokens. 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 2026-08-31.
Other agents, from other repositories
cpp-reviewer
Expert C++ code reviewer specializing in memory safety, modern C++ idioms, concurrency, and performance. Use for all C++ code changes. MUST BE USED for C++ projects.
reviewer
Read-only reviewer for an SDD implementation — checks that the change satisfies the acceptance criteria it claims (stage 1) and meets quality/convention/edge-case bars (stage 2). Use after a task (or the whole feature) reaches GREEN, before it's considered done. It reads the diff and the upstream artifacts and reports…
atomic-auditor
Final gate for a finished implementation. Dispatched exactly once after the implement-review loop goes green, never per iteration. Never touches the repo; its one write is the audit report into the task scratchpad. Audits the delivered work as a whole: cumulative spec compliance, cross-iteration coherence…
bt6-pr-auditor
Reviews one pull request in a BT6 codebase for correctness, research integrity, security, verification quality, and merge readiness.
Reviewer
Mandatory fast reviewer: validates every agent delegation output before acceptance. Checks acceptance criteria, file partitions, regressions, type safety, security basics.
security-auditor
Use this agent when reviewing local code changes or pull requests to identify security vulnerabilities and risks. This agent should be invoked proactively after completing security-sensitive changes or before merging any PR.