← ignarucli@0.3.0
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. 78 · Caution
Evidence
The sandbox observed nothing beyond the static findings; the proposal cites those.
Diff · removes 13 caught lines
Download .patchdiff --git a/harness/core/RULES.md b/harness/core/RULES.md --- a/harness/core/RULES.md +++ b/harness/core/RULES.md @@ -149,1 +149,0 @@ -- `image gen "<prompt>" --out <path>` / `image photo "<query>"`: generated art or a licence-safe photo. diff --git a/harness/skills/node-scene/SKILL.md b/harness/skills/node-scene/SKILL.md --- a/harness/skills/node-scene/SKILL.md +++ b/harness/skills/node-scene/SKILL.md @@ -84,1 +84,0 @@ -- `image gen "<prompt>" --out assets/texture.webp` for a texture or matcap; never for UI. diff --git a/README.md b/README.md --- a/README.md +++ b/README.md @@ -178,1 +178,0 @@ -ignaru ask "<prompt>" [--thread <id>] [--model <id>] Ask the agent @@ -180,1 +179,0 @@ -ignaru video "<prompt>" [--seconds <n>] [--aspect <a>] [--dry-run] diff --git a/harness/kit/cursor-effects.html b/harness/kit/cursor-effects.html --- a/harness/kit/cursor-effects.html +++ b/harness/kit/cursor-effects.html @@ -7,1 +7,0 @@ -kit: cursor-effects What: cursor-effects 1.0.18 followingDotCursor: a small dot that trails the pointer. One effect only. When: a single playful or editorial to diff --git a/harness/skills/design/SKILL.md b/harness/skills/design/SKILL.md --- a/harness/skills/design/SKILL.md +++ b/harness/skills/design/SKILL.md @@ -72,1 +72,0 @@ -| `image` | Generate an image, or fetch a licence-safe photo | `image gen "<prompt>" --out assets/x.webp` · `image photo "<query>"` | @@ -126,1 +125,0 @@ -5. When the person approves or rejects a direction, or corrects a choice, append one line to `.ignaru/memory.md`: diff --git a/harness/fonts/fonts.json b/harness/fonts/fonts.json --- a/harness/fonts/fonts.json +++ b/harness/fonts/fonts.json @@ -23,1 +23,0 @@ -"css": "https://fonts.googleapis.com/css2?family=Inter:wght@400;500;600;700&display=swap", @@ -63,1 +62,0 @@ -"css": "https://fonts.googleapis.com/css2?family=Geist:wght@300;400;500;600;700&display=swap", @@ -100,1 +98,0 @@ -"css": "https://fonts.googleapis.com/css2?family=Instrument+Sans:wght@400;500;600;700&display=swap", @@ -139,1 +136,0 @@ -"css": "https://fonts.googleapis.com/css2?family=Manrope:wght@400;500;600;700;800&display=swap", @@ -178,1 +174,0 @@ -"css": "https://fonts.googleapis.com/css2?family=Hanken+Grotesk:wght@400;500;600;700&display=swap", diff --git a/harness/kit/auto-animate.html b/harness/kit/auto-animate.html --- a/harness/kit/auto-animate.html +++ b/harness/kit/auto-animate.html @@ -7,1 +7,0 @@ -kit: auto-animate What: @formkit/auto-animate 0.10.0 on a plain list. Items that are added, removed or reordered animate on their own, with no keyframes. When:
Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 0.3.0. 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: ignarucli: Fake instruction boundary tag (KillerSkills verdict caution 78)
## What the honeypot observed KillerSkills ran **[email protected]** in an isolated sandbox (decoy credentials, recording proxy, syscall trace; engine 1.7) and scored it **CAUTION 78/100**. Summary: no network, touches nothing outside the project, writes no memory. - See the findings below. Full trace and findings: https://killerskills.ai/skills/npm/ignarucli/0.3.0 ## Findings with a location - `harness/core/RULES.md:149` · Fake instruction boundary tag — harness/core/RULES.md:149 · Fake instruction boundary tag · informational: not an instruction file - `harness/skills/node-scene/SKILL.md:84` · Fake instruction boundary tag — harness/skills/node-scene/SKILL.md:84 · Fake instruction boundary tag - `README.md:180` · Fake instruction boundary tag — README.md:180 · Fake instruction boundary tag · informational: not an instruction file - `harness/kit/cursor-effects.html:7` · HTML comment carrying imperative instructions to the agent — harness/kit/cursor-effects.html:7 · HTML comment carrying imperative instructions to the agent · informational: not an instruction file - `harness/skills/design/SKILL.md:126` · Instructs the agent to modify its persistent memory or instruction files — harness/skills/design/SKILL.md:126 · Instructs the agent to modify its persistent memory or instruction files - `harness/fonts/fonts.json:23` · URL with embedded credentials — harness/fonts/fonts.json:23 · URL with embedded credentials — may be phishing or credential leaking - `harness/fonts/fonts.json:63` · URL with embedded credentials — harness/fonts/fonts.json:63 · URL with embedded credentials — may be phishing or credential leaking - `harness/fonts/fonts.json:178` · URL with embedded credentials — harness/fonts/fonts.json:178 · URL with embedded credentials — may be phishing or credential leaking - `harness/fonts/fonts.json:100` · URL with embedded credentials — harness/fonts/fonts.json:100 · URL with embedded credentials — may be phishing or credential leaking - `harness/kit/auto-animate.html:7` · HTML comment carrying imperative instructions to the agent — harness/kit/auto-animate.html:7 · HTML comment carrying imperative instructions to the agent · informational: not an instruction file - `harness/fonts/fonts.json:139` · URL with embedded credentials — harness/fonts/fonts.json:139 · URL with embedded credentials — may be phishing or credential leaking - `README.md:178` · Fake instruction boundary tag — README.md:178 · Fake instruction boundary tag · informational: not an instruction file ## Proposed change The attached patch removes the 13 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._
# or from a terminal, under your own account: gh repo fork jawadjalal/ignaru --clone && cd ignaru curl -s https://killerskills.ai/skills/npm/ignarucli/0.3.0/patch.diff | git apply --unidiff-zero git checkout -b killerskills/0.3.0 && git commit -am "ignarucli: Fake instruction boundary tag" gh pr create --fill --body-file <(curl -s https://killerskills.ai/skills/npm/ignarucli/0.3.0/patch.md)
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.