← kane-cli-skill@0.1.6

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 16 caught lines

Download .patch
diff --git a/skills/references/explore.md b/skills/references/explore.md
--- a/skills/references/explore.md
+++ b/skills/references/explore.md
@@ -1,1 +1,0 @@
-kane-cli skill reference: exploration on HyperExecute, and the tests it designs into a shared project. Read when the person asks to explore an app, map its flow
diff --git a/skills/references/debug.md b/skills/references/debug.md
--- a/skills/references/debug.md
+++ b/skills/references/debug.md
@@ -1,1 +1,0 @@
-Read this when a kane-cli run failed and you need to diagnose. Owns the evidence-pack layout (the only source of run logs), how to read a sealed pack, the debug
diff --git a/skills/references/live-strip.md b/skills/references/live-strip.md
--- a/skills/references/live-strip.md
+++ b/skills/references/live-strip.md
@@ -1,1 +1,0 @@
-Read this when the person asks to watch runs live, asks about the status line, or right after a first run in Claude Code (to offer the strip). Owns what the liv
diff --git a/skills/references/policy.md b/skills/references/policy.md
--- a/skills/references/policy.md
+++ b/skills/references/policy.md
@@ -1,1 +1,0 @@
-kane-cli skill reference: evidence policies and verdicts (judge a sealed pack against a policy YAML: lint-policy, evidence verdict, --policy/--policy-id on test
diff --git a/skills/references/cards.md b/skills/references/cards.md
--- a/skills/references/cards.md
+++ b/skills/references/cards.md
@@ -1,1 +1,0 @@
-Read this when presenting the result of any kane-cli run, saved test or suite. Owns every result card: run passed, run failed, didn't start, stopped early, prod
diff --git a/skills/references/assurance-parsing.md b/skills/references/assurance-parsing.md
--- a/skills/references/assurance-parsing.md
+++ b/skills/references/assurance-parsing.md
@@ -1,1 +1,0 @@
-kane-cli skill reference: NDJSON wire contract for the assurance conversational commands (context ingest/extract / design tests / maintain reconcile / cover wit
diff --git a/skills/references/ready-check.md b/skills/references/ready-check.md
--- a/skills/references/ready-check.md
+++ b/skills/references/ready-check.md
@@ -1,1 +1,0 @@
-Read this at the start of every session that will use kane-cli. Owns the preflight script, the ready check card (full, one-line and problem versions), which pro
diff --git a/skills/SKILL.md b/skills/SKILL.md
--- a/skills/SKILL.md
+++ b/skills/SKILL.md
@@ -195,1 +195,0 @@
-…li the default for this project or their team ("always use kane for E2E here", "add kane to CLAUDE.md / AGENTS.md") → Read `references/setup-and-config.md` §Make kane-cli the default in a proje…
diff --git a/skills/references/first-run.md b/skills/references/first-run.md
--- a/skills/references/first-run.md
+++ b/skills/references/first-run.md
@@ -1,1 +1,0 @@
-Read this when the agent config shows no completed onboarding (a first session). Owns the run-first rule and its detected defaults, the first-run tour (verbatim
diff --git a/skills/references/agent-contract.md b/skills/references/agent-contract.md
--- a/skills/references/agent-contract.md
+++ b/skills/references/agent-contract.md
@@ -1,1 +1,0 @@
-kane-cli skill reference: the agent contract for a folder linked to a shared project (kane-cli 0.10.0+). Which agent flag each command takes, the NDJSON shapes,
diff --git a/skills/references/setup-and-config.md b/skills/references/setup-and-config.md
--- a/skills/references/setup-and-config.md
+++ b/skills/references/setup-and-config.md
@@ -107,1 +107,0 @@
-…he person asks** ("always use kane for E2E here", "set this repo up for kane", "add kane to CLAUDE.md / AGENTS.md"). Never do it during install, a first session, or a ready check, and never add…
@@ -115,1 +114,0 @@
-…oject (a home folder, a scratch folder): say so in one sentence and stop. Never write to a global file (`~/.claude/CLAUDE.md`, `~/.codex/AGENTS.md`): a global line would apply to every project the person has. **2. Look at wha…
diff --git a/skills/references/bugs.md b/skills/references/bugs.md
--- a/skills/references/bugs.md
+++ b/skills/references/bugs.md
@@ -1,1 +1,0 @@
-kane-cli skill reference: bug memory and feedback (kane-cli bugs list|show|mark|note|duplicate|retire|categories). Read it after a saved test failed and you are
diff --git a/skills/references/testmd.md b/skills/references/testmd.md
--- a/skills/references/testmd.md
+++ b/skills/references/testmd.md
@@ -1,1 +1,0 @@
-Read this when the user wants to save, persist, version-control, or re-run a browser test. Owns _test.md file format, replay/cascade rules, testmd subcommands, 
diff --git a/skills/references/objectives-cookbook.md b/skills/references/objectives-cookbook.md
--- a/skills/references/objectives-cookbook.md
+++ b/skills/references/objectives-cookbook.md
@@ -1,1 +1,0 @@
-Read this when authoring objective strings — for either `kane-cli run "..."` or step bodies inside a `_test.md`. Owns the full catalog of action verbs, assertio
diff --git a/skills/references/test-manager.md b/skills/references/test-manager.md
--- a/skills/references/test-manager.md
+++ b/skills/references/test-manager.md
@@ -1,1 +1,0 @@
-Read this for the Test Manager (TMS) agent surface — what gets uploaded, where the run lands, the `projects` / `folders` subcommands and their NDJSON wire shape

Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 0.1.6. 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: kane-cli-skill: HTML comment carrying imperative instructions to the agent (KillerSkills verdict caution 78)

## What the honeypot observed
KillerSkills ran **[email protected]** in an isolated sandbox (decoy credentials, recording proxy, syscall trace; engine 1.9.2) 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/kane-cli-skill/0.1.6

## Findings with a location
- `skills/references/explore.md:1` · HTML comment carrying imperative instructions to the agent — skills/references/explore.md:1 · HTML comment carrying imperative instructions to the agent
- `skills/references/debug.md:1` · HTML comment carrying imperative instructions to the agent — skills/references/debug.md:1 · HTML comment carrying imperative instructions to the agent
- `skills/references/live-strip.md:1` · HTML comment carrying imperative instructions to the agent — skills/references/live-strip.md:1 · HTML comment carrying imperative instructions to the agent
- `skills/references/policy.md:1` · HTML comment carrying imperative instructions to the agent — skills/references/policy.md:1 · HTML comment carrying imperative instructions to the agent
- `skills/references/cards.md:1` · HTML comment carrying imperative instructions to the agent — skills/references/cards.md:1 · HTML comment carrying imperative instructions to the agent
- `skills/references/assurance-parsing.md:1` · HTML comment carrying imperative instructions to the agent — skills/references/assurance-parsing.md:1 · HTML comment carrying imperative instructions to the agent
- `skills/references/ready-check.md:1` · HTML comment carrying imperative instructions to the agent — skills/references/ready-check.md:1 · HTML comment carrying imperative instructions to the agent
- `skills/SKILL.md:195` · Instructs the agent to modify its persistent memory or instruction files — skills/SKILL.md:195 · Instructs the agent to modify its persistent memory or instruction files
- `skills/references/first-run.md:1` · HTML comment carrying imperative instructions to the agent — skills/references/first-run.md:1 · HTML comment carrying imperative instructions to the agent
- `skills/references/agent-contract.md:1` · HTML comment carrying imperative instructions to the agent — skills/references/agent-contract.md:1 · HTML comment carrying imperative instructions to the agent
- `skills/references/setup-and-config.md:107` · Instructs the agent to modify its persistent memory or instruction files — skills/references/setup-and-config.md:107 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `skills/references/setup-and-config.md:115` · Instructs the agent to modify its persistent memory or instruction files — skills/references/setup-and-config.md:115 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file

## Proposed change
The attached patch removes the 16 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 LambdaTest/kane-cli --clone && cd kane-cli
curl -s https://killerskills.ai/skills/npm/kane-cli-skill/0.1.6/patch.diff | git apply --unidiff-zero
git checkout -b killerskills/0.1.6 && git commit -am "kane-cli-skill: HTML comment carrying imperative instructions to the agent"
gh pr create --fill --body-file <(curl -s https://killerskills.ai/skills/npm/kane-cli-skill/0.1.6/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.