diff --git a/corpus/CLAUDE.learned.md b/corpus/CLAUDE.learned.md
index c2753870..29611a5a 100644
--- a/corpus/CLAUDE.learned.md
+++ b/corpus/CLAUDE.learned.md
@@ -42,3 +42,4 @@ Engine-only install drops this file; reflect Accepted global rules land here.
 - When handing off a blocked action for me to run myself, always write it to a small script file and hand back exactly one `! bash <path>` / `! python3 <path>` line — never paste inline multi-line, multi-flag, or `&&`/`&`-backgrounded shell text into chat for me to copy. This applies on the very first handoff and every one after it, not after I complain about copy-paste pain. Pasted shell breaks in transit — wrapped lines, lost quoting, an `&` that detaches a job — while one `!` line runs exactly the file that was written.
 - Never pass a quoted command string through `sudo -i`, `su -`, or `su -l`. Those start a login shell that reads the command a second time, after the first shell already used up the quotes, so the string splits on spaces and its first word runs alone (`sudo -u demo -H -i bash -lc 'set -u; echo one'` fails with `set: -c: invalid option`; the same line without `-i` prints `one`). Copy the script to the host and run it by path, or drop `-i`.
 - The shell behind the Bash tool may be zsh, which does not split an unquoted variable into words: `for x in $LIST` runs once with the whole list as one item, exits 0, and reads as "all done". Run any loop over a list under bash (`bash <<'EOF' ... EOF`), count lookups that failed as unchecked rather than done, and spot-check one item before acting on the loop's summary. This is zsh's `SH_WORD_SPLIT` option being off by default (zsh manual, Options, https://zsh.sourceforge.io/Doc/Release/Options.html#index-SH_005fWORD_005fSPLIT).
+- Before asking me whether to run a script, read its whole body, not its header comment, usage text, or `--dry-run` output, and name in the question every process it kills, file it deletes, and service it restarts. A header states what the author meant; the body is what runs, so a consent question built from the header asks me to approve something other than what will happen. No known prior art.
diff --git a/tests/test_read_script_before_consent_rule.py b/tests/test_read_script_before_consent_rule.py
new file mode 100644
index 00000000..c4b3ebe6
--- /dev/null
+++ b/tests/test_read_script_before_consent_rule.py
@@ -0,0 +1,45 @@
+#!/usr/bin/env python3
+"""Pin the learned rule: read a script's whole body before asking to run it.
+
+A detector cannot judge this reliably, so the test proves the rule is present
+with its trigger, required action, and reason; deleting or hollowing it fails here.
+"""
+import os
+import unittest
+
+REPO_ROOT = os.path.dirname(os.path.dirname(os.path.abspath(__file__)))
+LEARNED = os.path.join(REPO_ROOT, "corpus", "CLAUDE.learned.md")
+
+
+def working_style_section():
+    with open(LEARNED, encoding="utf-8") as handle:
+        text = handle.read()
+    start = text.index("# Working style")
+    end = text.find("\n# ", start + 1)
+    return text[start:] if end == -1 else text[start:end]
+
+
+def bullet_starting(prefix):
+    for line in working_style_section().splitlines():
+        if line.startswith(prefix):
+            return line
+    raise AssertionError(f"no working-style rule starts with {prefix!r}")
+
+
+class TestReadScriptBeforeConsent(unittest.TestCase):
+    PREFIX = "- Before asking me whether to run a script"
+
+    def test_rule_requires_the_whole_body_not_the_header(self):
+        rule = bullet_starting(self.PREFIX)
+        self.assertIn("read its whole body", rule)
+        self.assertIn("header comment", rule)
+        self.assertIn("`--dry-run` output", rule)
+
+    def test_rule_names_what_the_question_must_list(self):
+        rule = bullet_starting(self.PREFIX)
+        for effect in ("process it kills", "file it deletes", "service it restarts"):
+            self.assertIn(effect, rule)
+
+
+if __name__ == "__main__":
+    unittest.main()