Propose a patch

A fix the maintainer can act on, with the evidence attached. It opens under your GitHub account in your browser; KillerSkills holds no token and never modifies the package. The fix is verified the only way we verify anything: the next crawl re-runs the new version in the honeypot. 65 · Caution

Evidence

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

Diff · removes 6 caught lines

Download .patch

diff --git a/references/troubleshooting.md b/references/troubleshooting.md
--- a/references/troubleshooting.md
+++ b/references/troubleshooting.md
@@ -62,1 +62,0 @@
-python3 "$IMAGE_FORGE_SCRIPT" edit --brief '<approved brief>' --image '<path>' --prompt '<instruction>'
diff --git a/references/openai-compatible-images.md b/references/openai-compatible-images.md
--- a/references/openai-compatible-images.md
+++ b/references/openai-compatible-images.md
@@ -27,1 +27,0 @@
-python3 "$IMAGE_FORGE_SCRIPT" generate --brief '<approved brief>' --prompt '<prompt>'
@@ -37,1 +36,0 @@
-"prompt": "<prompt>",
@@ -56,1 +54,0 @@
-python3 "$IMAGE_FORGE_SCRIPT" edit --brief '<approved brief>' --image '<path>' --prompt '<instruction>'
diff --git a/references/installation.md b/references/installation.md
--- a/references/installation.md
+++ b/references/installation.md
@@ -93,1 +93,0 @@
-python3 "$IMAGE_FORGE_SCRIPT" generate --brief '<approved brief>' --prompt '<prompt>'
@@ -99,1 +98,0 @@
-python3 "$IMAGE_FORGE_SCRIPT" edit --brief '<approved brief>' --image '<path>' --prompt '<instruction>'

Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 0.1.1. 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: nextai-image-forge: Fake instruction boundary tag (KillerSkills verdict caution 65)

## 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 65/100**. Summary: static analysis only — behavior not yet observed.

- See the findings below.

Full trace and findings: https://killerskills.ai/skills/clawhub/nextai-image-forge/0.1.1

## Findings with a location
- `references/troubleshooting.md:62` · Fake instruction boundary tag — references/troubleshooting.md:62 · Fake instruction boundary tag · informational: not an instruction file
- `references/openai-compatible-images.md:56` · Fake instruction boundary tag — references/openai-compatible-images.md:56 · Fake instruction boundary tag · informational: not an instruction file
- `references/openai-compatible-images.md:37` · Fake instruction boundary tag — references/openai-compatible-images.md:37 · Fake instruction boundary tag · informational: not an instruction file
- `references/installation.md:99` · Fake instruction boundary tag — references/installation.md:99 · Fake instruction boundary tag · informational: not an instruction file
- `references/openai-compatible-images.md:27` · Fake instruction boundary tag — references/openai-compatible-images.md:27 · Fake instruction boundary tag · informational: not an instruction file
- `references/installation.md:93` · Fake instruction boundary tag — references/installation.md:93 · Fake instruction boundary tag · informational: not an instruction file

## Proposed change
The attached patch removes the 6 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._

No GitHub source is recorded for this package, so there is nothing to open; paste the body into the registry's support channel.

We count proposals opened, not merged: a merge shows up as a new version that the crawl re-runs. Disputes: the maintainer can contest the verdict from its page; a person reviews it and the outcome is published.