← Back to AI Insights
Gemini Executive Synthesis

LLM-driven documentation generation in Simplified Technical English (STE), specifically the 'check mode' functionality for reporting rule violations.

Technical Positioning
Accurate and verifiable citation of STE rule numbers, preventing LLM hallucination and ensuring auditability of compliance reports.
SaaS Insight & Market Implications
This issue reveals a critical internal inconsistency within the SimpleEnglish agent's documentation, directly impacting its 'check mode' functionality. The `checklist.md` file incorrectly directs agents to a non-existent `rules.md` for rule citations, while the actual rule catalog resides in `SKILL.md`. This architectural flaw forces LLMs to either omit citations or, worse, hallucinate rule numbers, precisely the failure `SKILL.md` explicitly warns against. For developers adopting this skill in production environments, particularly for full audits, this undermines the integrity and verifiability of compliance reports. The market demands automated tools that provide accurate, auditable references for rule violations. This defect compromises the tool's reliability and increases the operational burden of verifying LLM output, directly contradicting the 'no AI slop' promise.
Proprietary Technical Taxonomy
check mode rule numbers rule catalog LLM hallucination monorepo citations SKILL.md checklist.md

Raw Developer Origin & Technical Request

Source Icon GitHub Issue Jul 31, 2026
Repo: AminBlg/SimpleEnglish
checklist.md tells check mode to cite rule numbers from a rules.md that does not exist

`references/checklist.md` points check mode at a `rules.md` that the package does not contain.

**Where**

`skills/simple-english/references/checklist.md`, in the "When reporting violations (check mode)" section:

> For each violation give: the rule number, the offending text, and a compliant rewrite. Cite only rule numbers that appear in rules.md.

**The problem**

The package ships `SKILL.md`, `references/checklist.md` and `references/use-cases.md`. There is no `rules.md`. The rule catalog lives in `SKILL.md`.

`SKILL.md` states the opposite, and explains why it matters:

> Cite only rule numbers that exist in this file. Do not cite rule numbers from memory: the numbering is unintuitive and models invent it (tested — an agent without this file cited "Rule 3.1: short sentences"; the real Rule 3.1 is about verb forms).

So an agent that follows the checklist is sent to a file it cannot read. It then either omits the citations or invents them, which is the failure `SKILL.md` warns about. `checklist.md` is also the path `SKILL.md` recommends for a full audit, so this is on the main route for check mode, not an edge case.

**Suggested fix**

Change the reference in `checklist.md` from `rules.md` to `SKILL.md`. Alternatively, split the catalog out into a real `rules.md` and have `SKILL.md` reference it.

**Context**

Found while adopting the skill in a monorepo and running it in check mode across seven documents. The citations came out correct only because `SKILL.md` was in conte...

Developer Debate & Comments

No active discussions extracted for this entry yet.

Adjacent Repository Pain Points

Other highly discussed features and pain points extracted from AminBlg/SimpleEnglish.

Extracted Positioning
The `npx skills add` command for installing agent skills, specifically its telemetry functionality.
User consent and data privacy in developer tooling. The issue implies a breach of trust regarding 'unauthorized connection' and 'exfiltrated private data'.
Top Replies
sneak • Jul 31, 2026
https://github.com/vercel-labs/skills/blob/7cb7db64dc1201052dea305e508a2fc490f7e5e2/src/telemetry.ts#L1
emilsoman • Jul 31, 2026
If you install skills using "npx skills", it sends telemetry events to skills.sh for ranking skills. See https://www.skills.sh/docs . You can turn it off using a cli flag
sneak • Jul 31, 2026
Yes, I am aware of this - now - only after it has exfiltrated private data from my machine without consent.
Extracted Positioning
LLM-driven documentation generation in Simplified Technical English (STE), specifically the 'vocabulary discipline table' for part-of-speech rulings.
Accurate adherence to ASD-STE100 Issue 9 dictionary, ensuring precise linguistic control for technical documentation and factual correctness.
Extracted Positioning
LLM-driven documentation generation in Simplified Technical English (STE), specifically the 'consistency pass' for vocabulary discipline.
Strict adherence to ASD-STE100 Issue 9 dictionary for technical English compliance, aiming for 'no AI slop' and high-quality, rule-based output.

Frequently Asked Questions

Market intelligence mapped to LLM-driven documentation generation in Simplified Technical English (STE), specifically the 'check mode' functionality for reporting rule violations..

What problem does LLM-driven documentation generation in Simplified Technical English (STE), specifically the 'check mode' functionality for reporting rule violations. solve?
Based on our AI analysis of the original developer request, its primary technical positioning is: Accurate and verifiable citation of STE rule numbers, preventing LLM hallucination and ensuring auditability of compliance reports.
What are the foundational technologies related to LLM-driven documentation generation in Simplified Technical English (STE), specifically the 'check mode' functionality for reporting rule violations.?
Our proprietary extraction maps LLM-driven documentation generation in Simplified Technical English (STE), specifically the 'check mode' functionality for reporting rule violations. to adjacent architectural concepts including check mode, rule numbers, rule catalog, LLM hallucination.

Engagement Signals

0
Replies
open
Issue Status

Cross-Market Term Frequency

Quantifies the cross-market adoption of foundational terms like audit and SKILL.md by tracking occurrence frequency across active SaaS architectures and enterprise developer debates.