archi-silent-audit

archi-silent-audit is a skill for Claude Code from JiuNian3219/architext. It costs 37 tokens per session (2,125 once invoked), scanned A, original, MIT.

A protocol-only helper for reviewing code and documents against defined review dimensions. It reports findings only when it can provide the original evidence and its location.

In plain words
What is it for?
Use it in supported architecture workflows to inspect plans or code and return evidence-backed findings. It does not decide how those findings should be fixed.
Why use it?
It prevents unsupported guesses from becoming review findings and ensures the requested review dimensions are checked together.

Skill for Claude Code

Written for Claude Code: disable-model-invocation in frontmatter. Also seen: mentions subagents.

Good fit Use it in supported architecture workflows to inspect plans or code and return evidence-backed findings. It does not decide how those findings should be fixed.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/jiunian3219/architext/archi-silent-audit
Install

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.

Any agent
npx skills add JiuNian3219/architext --skill archi-silent-audit
Clone the repo
git clone --depth 1 https://github.com/JiuNian3219/architext

Made for: Claude Code.

Wrote 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.

agentmods badge for archi-silent-audit

README.md
[![agentmods](https://agentmods.dev/badge/skills/jiunian3219/architext/archi-silent-audit/github.svg)](https://agentmods.dev/skills/jiunian3219/architext/archi-silent-audit)
Your own site
<a href="https://agentmods.dev/skills/jiunian3219/architext/archi-silent-audit"><img src="https://agentmods.dev/badge/skills/jiunian3219/architext/archi-silent-audit/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.

agentmods 80×15 button for archi-silent-audit

Your own site · 80×15
<a href="https://agentmods.dev/skills/jiunian3219/architext/archi-silent-audit"><img src="https://agentmods.dev/badge/skills/jiunian3219/architext/archi-silent-audit.svg" alt="Reviewed on agentmods" width="80" height="20"></a>
Per session 37 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 2,125 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5.1 $0.00037 $0.02125
Opus 5 $0.00018 $0.01063
Sonnet 5 $0.00007 $0.00425
Haiku 4.5 $0.00004 $0.00213

Measured 11d ago against content hash 05af81d72d8f, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-11, from the pricing page.

Security

Grade A, and why

archi-silent-audit 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 11d 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.

templates/en/skills/archi-silent-audit/SKILL.md · 103 lines

How it starts

The opening of the file, as written. The whole thing — 103 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Invocation

  • Auto-invoke: No, not triggered by model based on description.
  • Trigger location: Only explicitly called via [[SUBAGENT]] / [[NO-SUBAGENT]] in /archi.* protocols.
  • Execution context: When subagent supported must execute in independent subagent/independent context; only downgrade to inline Skill when no subagent.
  • Boundary: Only return protocol-required structured artifacts, subsequent write, confirm and signoff handled by calling protocol.

Core Principles

  • No Evidence No Report: Each finding must have evidence (original snippet + location); guess without locatable evidence never reported
  • Pre-bound Level: severity defined by dimension library table below, skill only judges "Triggered / Not triggered" not level
  • All Open Once: Dimensions filtered by mode run all at once, not staged, not early-stop
  • Level ≠ Action: Findings and levels returned as-is, how to respond decided by calling protocol

Dimension Library

Symbol legend:

  • Mode column: I = init, P = plan-docs, C = code-impl (Multiple use / separator)
  • Gate column: [[WHEN: X]] = Deploy-time features filter; runtime: X = Judgment at call-time via task_meta; = No gate
# ID Name Mode Level Gate Trigger Condition
1 VISION_ROADMAP_ALIGN Vision-Roadmap Alignment I CRITICAL roadmap task goal direction inconsistent with vision.md north star direction (task doesn't serve north star)
2 TECH_STACK_DECLARED Tech Stack Declaration Consistent I / P / C CRITICAL init: tech_stack.md doesn't reflect package.json actual dependencies; plan-docs: spec/plan uses library/framework/API not declared in tech_stack; code-impl: code imports library not in tech_stack
3 GLOBAL_FILES_COMPLETE Global Files Completeness I CRITICAL vision / roadmap / map / dictionary / tech_stack / custom_rules any missing or only skeleton without substance
4 INFORMATION_ROUTING Information No Leakage I WARNING Information in Brief / code (feature points / terms / constraints) not routed to corresponding global files
5 DESIGN_TOKENS_BASIC Design Tokens Basics Ready I CRITICAL [[WHEN: ui]] design_tokens.json missing aestheticDirection.preset / primitivePalette.brand / core font size fields
6 DESIGN_FIDELITY Design Fidelity P CRITICAL spec § 2 doesn't cover feature design points confirmed in discussion (has omissions)
7 DIMENSION_MATCH Dimension Match P WARNING spec § 2 dimension format doesn't match Task Type (didn't output per Task Type specified dimensions)
8 WBS_COVERAGE WBS Coverage P CRITICAL plan.json has spec AC entries not covered by any task
9 NOTES_QUALITY Notes Quality P WARNING plan.json task.notes lacks "Output:" / "Constraint:" / "Verification:" any keyword; or "Verification:" value contains "appropriately", "reasonably", "as needed", "per situation" etc vague quantifiers
10 INTERFACE_EXPORTS Interface Exports P CRITICAL INF task § 4 not filled; or downstream deps exist but interface signature not declared
11 CONSTRAINTS_REDLINES Constraints Red Lines P WARNING spec § 5 lacks red lines related to vision.md north star or tech_stack known prohibited items
12 DATA_INTEGRITY_SPEC Data Integrity (spec) P CRITICAL [[WHEN: data]] spec entity/field inconsistent with data_snapshot.json existing core entity/field
13 DESIGN_TRACE Design Trace P WARNING runtime: is_complex design.md § 6 trace table has AC not mapped to design node
14 PARAMETER_SPECIFICITY Parameter Specificity P INFO runtime: is_complex design.md § 3 parameter table has "appropriately", "reasonably", "as needed", "per situation", "some" etc vague quantifiers
15 DESIGN_SELF_CHECK Design Self-Check P CRITICAL runtime: is_complex design.md § 2 mechanism self-check list any item marked ✗ or empty unchecked
15.5 TEST_PLAN_QUALITY Test Plan Quality P WARNING plan.json tests / task notes only write "add tests", "run tests", "cover logic" etc placeholder sentences; or not mapped to spec AC / bug reproduction steps
16 TECH_STACK_STYLE_MATCH Tech Stack Style Consistent C WARNING Code pattern / API style deviates from tech_stack.md declaration (e.g. declared ESM but uses CommonJS, declared hooks-first but uses class component)
17 SOTA SOTA C WARNING Appears in tech_stack explicitly listed anti-pattern; or uses deprecated API replaced by same stack
18 SECURITY Security C CRITICAL Code has hardcoded secrets / plaintext passwords; or external input unvalidated directly into DB / FS / shell / eval
19 PERFORMANCE Performance C WARNING Obvious full imports, repeated computation, unreleased listeners, O(n²) optimizable to O(n) scenarios
20 UI_TOKEN_COMPLIANCE UI Token Compliance C CRITICAL [[WHEN: ui]] + runtime: involves_ui Style values have hardcoded color / font size / spacing (not var(--*) or token reference)
21 ACCESSIBILITY Accessibility C WARNING [[WHEN: ui]] + runtime: involves_ui Interactive elements lack aria-label / semantic tags / keyboard reachability (button / input / link missing required attributes)
22 DATA_INTEGRITY_CODE Data Integrity (code) C CRITICAL [[WHEN: data]] + runtime: involves_data Code implemented field name / type inconsistent with data_snapshot.json
23 DESIGN_COMPLIANCE_STATE Design Compliance State C CRITICAL runtime: is_complex Code state transition / flow / protocol inconsistent with design.md § 2
24 INVARIANT_ENFORCEMENT Invariant Enforcement C CRITICAL runtime: is_complex design.md § 4 invariants not enforced in code via assert / runtime check / type guard
25 PARAMETER_ALIGNMENT Parameter Alignment C CRITICAL runtime: is_complex Code values / thresholds inconsistent with design.md § 3 parameter table
26 TEST_BEHAVIOR_QUALITY Test Behavior Quality C WARNING New/modified tests only assert mock call counts, file exists, object non-empty, JSON parseable; lacking behavior results, boundary/error paths or Red→Green reproduction evidence

Read the full file on GitHub · 103 lines

Changes

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.

  1. 11d ago First seen · 103 lines · 37 tokens per session scan A 05af81d72d8f

Subscribe to this mod's changes

archi-silent-audit is a skill published in the GitHub repository JiuNian3219/architext (14 stars, last pushed 3mo ago), licensed MIT. It adds 37 tokens to every session and 2,125 once invoked, about $0.0002 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.