Propose a patch

Evidence

The sandbox observed nothing beyond the static findings; the proposal cites those.

diff --git a/references/requirement-writing-standards.md b/references/requirement-writing-standards.md
--- a/references/requirement-writing-standards.md
+++ b/references/requirement-writing-standards.md
@@ -35,1 +35,0 @@
-| Ubiquitous | The `<system>` shall `<response>` | The system shall encrypt all data at rest. |
@@ -36,1 +35,0 @@
-| Event-driven | WHEN `<trigger>`, the `<system>` shall `<response>` | WHEN a user submits an invalid email format, the system shall display an inline validation error. |
@@ -37,1 +35,0 @@
-| State-driven | WHILE `<state>`, the `<system>` shall `<response>` | WHILE the device is offline, the system shall queue write operations locally. |
@@ -38,1 +35,0 @@
-| Unwanted behavior | IF `<condition>`, THEN the `<system>` shall `<response>` | IF the payment gateway times out, THEN the system shall retry up to 3 times before surfacing a failure state. |
@@ -39,1 +35,0 @@
-…ing a failure state. | | Optional feature | WHERE `<feature is included>`, the `<system>` shall `<response>` | WHERE push notifications are enabled, the system shall re…

Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 1.0.2. If the behavior is intended, declare it (host under permissions.network, a line in the README) instead of removing it; the next crawl scores declarations.

Issue / PR body

Title: prd-creator: Fake instruction boundary tag (KillerSkills verdict caution 71)

## What the honeypot observed
KillerSkills ran **[email protected]** in an isolated sandbox (decoy credentials, recording proxy, syscall trace; engine 1.8.1) and scored it **CAUTION 71/100**. Summary: static analysis only — behavior not yet observed.

- See the findings below.

Full trace and findings: https://killerskills.ai/skills/clawhub/prd-creator/1.0.2

## Findings with a location
- `references/requirement-writing-standards.md:36` · Fake instruction boundary tag — references/requirement-writing-standards.md:36 · Fake instruction boundary tag · informational: not an instruction file
- `references/requirement-writing-standards.md:38` · Fake instruction boundary tag — references/requirement-writing-standards.md:38 · Fake instruction boundary tag · informational: not an instruction file
- `references/requirement-writing-standards.md:39` · Fake instruction boundary tag — references/requirement-writing-standards.md:39 · Fake instruction boundary tag · informational: not an instruction file
- `references/requirement-writing-standards.md:37` · Fake instruction boundary tag — references/requirement-writing-standards.md:37 · Fake instruction boundary tag · informational: not an instruction file
- `references/requirement-writing-standards.md:35` · Fake instruction boundary tag — references/requirement-writing-standards.md:35 · Fake instruction boundary tag · informational: not an instruction file

## Proposed change
The attached patch removes the 5 lines above. If the behavior is intended, the fix is to declare it instead: list the host under `permissions.network` (or the MCP manifest) and say in the README why the package needs it; the next crawl will re-score it with the declaration taken into account.

## How this will be verified
Nothing is "verified safe" by assertion. When a new version is published, KillerSkills re-runs it in the honeypot and the verdict page updates; you can trigger it early from the verdict page ("Re-scan after I publish a fix"). Reply here or at [email protected] if you believe the verdict is wrong; disputes are reviewed by a person and the outcome is published on the verdict page.

_Proposed through killerskills.ai/threats by a user of the package, under their own GitHub account. The evidence is public; the package was never modified by KillerSkills._