Propose a patch

Evidence

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

diff --git a/higher-ed.md b/higher-ed.md
--- a/higher-ed.md
+++ b/higher-ed.md
@@ -102,1 +102,0 @@
-…t run better rather than merely different. **After every supervision meeting**, add its dated line to `## Supervisions` in `~/Clawic/data/teacher/memory.md` the same day — what was agreed and what they will submit — with the milestone …
diff --git a/subjects.md b/subjects.md
--- a/subjects.md
+++ b/subjects.md
@@ -112,1 +112,0 @@
-…very year in every school. **When a wrong answer recurs in your own classes**, add it to `## Misconceptions` in `memory.md` with the question that catches it — the general list here is a starting point,…
diff --git a/explaining.md b/explaining.md
--- a/explaining.md
+++ b/explaining.md
@@ -121,1 +121,0 @@
-…n problems | **When an explanation, analogy or worked example finally lands**, add the row to `## Explanations That Landed` in `~/Clawic/data/teacher/memory.md`: topic, what worked, and why it worked. **When a wrong answer recurs across st…
diff --git a/classroom.md b/classroom.md
--- a/classroom.md
+++ b/classroom.md
@@ -139,1 +139,0 @@
-… finished. **When a behaviour approach works or fails with a specific group**, write it to `## What Works With This Group` in `memory.md` — technique, group, observed effect. **When a routine, seating rationale or be…
diff --git a/curriculum.md b/curriculum.md
--- a/curriculum.md
+++ b/curriculum.md
@@ -107,1 +107,0 @@
-…r/artifacts/<kebab-name>.md` with the date and the hours budget it assumes, and add its `## Boxes` line to `memory.md` in the same turn (`memory-template.md`). **When a coverage audit runs**, updat…
diff --git a/observation.md b/observation.md
--- a/observation.md
+++ b/observation.md
@@ -45,1 +45,0 @@
-…ose by agreeing the target, how it will be measured, and the review date — then write it to `## Practice Goals` in `memory.md` with the review in `## Due`. An improvement target with no review date is a co…
@@ -106,1 +105,0 @@
-…nce is already written. **After any observation, appraisal or peer feedback**, write the target to `## Practice Goals` in `~/Clawic/data/teacher/memory.md` — what was seen, the one change, how it will be measured — and put the review …
diff --git a/online.md b/online.md
--- a/online.md
+++ b/online.md
@@ -112,1 +112,0 @@
-…the platform, network, device baseline or timetable costs effort to find out**, write it to `## Environment` in `memory.md` — that is the file that stops the same twenty minutes being spent next term. *…
diff --git a/engagement.md b/engagement.md
--- a/engagement.md
+++ b/engagement.md
@@ -99,1 +99,0 @@
-…sked privately. **When a technique visibly moves participation with a group**, write it to `## What Works With This Group` in `memory.md`. **When a grouping or seating arrangement works or fails**, update `## Dynamic…
diff --git a/checking.md b/checking.md
--- a/checking.md
+++ b/checking.md
@@ -110,1 +110,0 @@
-… in the same lesson. **When a hinge question exposes a shared misconception**, add the row to `## Misconceptions` in `~/Clawic/data/teacher/memory.md`: topic, the wrong answer, and the question that catches it. **When an item wor…
diff --git a/planning.md b/planning.md
--- a/planning.md
+++ b/planning.md
@@ -114,1 +114,0 @@
-…gs that held, the part that ran late, and the misconception that appeared, then add its `## Boxes` line to `memory.md` in the same turn (`memory-template.md`). **When the lesson ran ahead or behind…

Zero-context hunks: apply with git apply --unidiff-zero fix.patch on the tagged source of 1.0.2. 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: teacher: Instructs the agent to modify its persistent memory or instruction files (KillerSkills verdict caution 72)

## 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 72/100**. Summary: static analysis only — behavior not yet observed.

- See the findings below.

Full trace and findings: https://killerskills.ai/skills/clawhub/teacher/1.0.2

## Findings with a location
- `higher-ed.md:102` · Instructs the agent to modify its persistent memory or instruction files — higher-ed.md:102 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `subjects.md:112` · Instructs the agent to modify its persistent memory or instruction files — subjects.md:112 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `explaining.md:121` · Instructs the agent to modify its persistent memory or instruction files — explaining.md:121 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `classroom.md:139` · Instructs the agent to modify its persistent memory or instruction files — classroom.md:139 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `curriculum.md:107` · Instructs the agent to modify its persistent memory or instruction files — curriculum.md:107 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `observation.md:106` · Instructs the agent to modify its persistent memory or instruction files — observation.md:106 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `online.md:112` · Instructs the agent to modify its persistent memory or instruction files — online.md:112 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `observation.md:45` · Instructs the agent to modify its persistent memory or instruction files — observation.md:45 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `engagement.md:99` · Instructs the agent to modify its persistent memory or instruction files — engagement.md:99 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `checking.md:110` · Instructs the agent to modify its persistent memory or instruction files — checking.md:110 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
- `planning.md:114` · Instructs the agent to modify its persistent memory or instruction files — planning.md:114 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file

## Proposed change
The attached patch removes the 11 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._