Propose a patch

Evidence

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

diff --git a/site/index.html b/site/index.html
--- a/site/index.html
+++ b/site/index.html
@@ -189,1 +189,0 @@
-Interactive Packer / Token Estimator
@@ -431,1 +430,0 @@
-<summary>How accurate is the token estimator?</summary> <p>ContextPack's tokenizer uses a calibrated heuristic calibrated against Claude and cl100k_base models.
@@ -436,1 +434,0 @@
-<summary>How does token budgeting handle large codebases?</summary> <p>When you set <code>--max-tokens</code>, ContextPack reserves tokens for header metadata a
@@ -441,1 +438,0 @@
-<summary>Does ContextPack send any code over the internet?</summary> <p>Never. All file operations, tree generations, token counts, and diffs are computed local
diff --git a/site/samples/contextpack/report.json b/site/samples/contextpack/report.json
--- a/site/samples/contextpack/report.json
+++ b/site/samples/contextpack/report.json
@@ -559,1 +559,0 @@
-"fix": "Add AGENTS.md if developers or coding agents integrate with your product."

Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 0.1.0. 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: contextpack-mcp: Collapsed <details> block carrying instructions (KillerSkills verdict caution 68)

## 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 68/100**. Summary: static analysis only — behavior not yet observed.

- See the findings below.

Full trace and findings: https://killerskills.ai/skills/glama/contextpack-mcp/0.1.0

## Findings with a location
- `site/index.html:436` · Collapsed <details> block carrying instructions — site/index.html:436 · Collapsed <details> block carrying instructions · informational: not an instruction file · informational: not an instruction file
- `site/index.html:431` · Collapsed <details> block carrying instructions — site/index.html:431 · Collapsed <details> block carrying instructions · informational: not an instruction file · informational: not an instruction file
- `site/samples/contextpack/report.json:559` · Instructs the agent to modify its persistent memory or instruction files — site/samples/contextpack/report.json:559 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file · informational: not an instruction file
- `site/index.html:189` · HTML comment carrying imperative instructions to the agent — site/index.html:189 · HTML comment carrying imperative instructions to the agent · informational: not an instruction file · informational: not an instruction file
- `site/index.html:441` · Collapsed <details> block carrying instructions — site/index.html:441 · Collapsed <details> block carrying instructions · informational: not an instruction file · 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._