Propose a patch

Evidence

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

diff --git a/evals/cases.json b/evals/cases.json
--- a/evals/cases.json
+++ b/evals/cases.json
@@ -63,1 +63,0 @@
-{ "id": "dedup-credentials-flagged", "category": "dossier-lint", "tool": "dedup_and_lint_dossier", "input": { "urls": ["https://user:pw@localhost/"] }, "expect": { "ok": false, "dedup": { "invalid": [
@@ -132,1 +131,0 @@
-{"id":"uni-anchor-leading-bom-rejected","category":"unicode-edge","tool":"evaluate_agent_response","input":{"response":"ANCHOR: a","contract":{"anchor":"a"}},"expect":{"ok":false,"mentions":["first l
@@ -133,1 +131,0 @@
-{"id":"uni-anchor-trailing-zwsp-rejected","category":"unicode-edge","tool":"evaluate_agent_response","input":{"response":"ANCHOR: a​","contract":{"anchor":"a"}},"expect":{"ok":false,"mentions":["first
@@ -136,1 +133,0 @@
-{"id":"uni-claim-zero-width-split-rejected","category":"unicode-edge","tool":"evaluate_agent_response","input":{"response":"ANCHOR: a\nThis is guar​ant‍eed.","contract":{"anchor":"a","forbiddenClaims"
@@ -137,1 +133,0 @@
-{"id":"uni-claim-soft-hyphen-rejected","category":"unicode-edge","tool":"evaluate_agent_response","input":{"response":"ANCHOR: a\nThis is guar­anteed.","contract":{"anchor":"a","forbiddenClaims":["gua
diff --git a/test/evaluate.test.mjs b/test/evaluate.test.mjs
--- a/test/evaluate.test.mjs
+++ b/test/evaluate.test.mjs
@@ -51,1 +51,0 @@
-assert.equal(foldForMatch('guar​an­teed'), 'guaranteed');

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: agent-eval-policy: Hidden zero-width characters detected (KillerSkills verdict caution 68)

## 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 68/100**. Summary: no network, touches nothing outside the project, writes no memory.

- See the findings below.

Full trace and findings: https://killerskills.ai/skills/glama/agent-eval-policy/0.3.0

## Findings with a location
- `evals/cases.json:133` · Hidden zero-width characters detected — evals/cases.json:133 · Hidden zero-width characters detected — may conceal content
- `evals/cases.json:63` · URL with embedded credentials — evals/cases.json:63 · URL with embedded credentials — may be phishing or credential leaking
- `test/evaluate.test.mjs:51` · Hidden zero-width characters detected — test/evaluate.test.mjs:51 · Hidden zero-width characters detected — may conceal content
- `evals/cases.json:136` · Hidden zero-width characters detected — evals/cases.json:136 · Hidden zero-width characters detected — may conceal content
- `evals/cases.json:132` · Hidden zero-width characters detected — evals/cases.json:132 · Hidden zero-width characters detected — may conceal content
- `evals/cases.json:137` · Hidden zero-width characters detected — evals/cases.json:137 · Hidden zero-width characters detected — may conceal content

## 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._