What the reviewer found
Prompting-research doc for a scaffolding agent, describing defenses against untrusted lifeboat inputs. Every flagged string ("IGNORE PREVIOUS INSTRUCTIONS... /etc/passwd", the HTML-comment injection, the ~/.ssh/authorized_keys path traversal) is an explicit labeled test canary or attack example the schema validator is built to reject, not a live instruction.
example-or-fixture— an example or test fixture
What was read
The file as it ships in intentdriven/abcd:
.abcd/development/research/prompting/agents/embark-scaffolder.md
What the static scan said
The scan flagged 3things. The reviewer kept 0 and dismissed 3 as false.
P1Instruction-override phrasing — false positiveP2Hidden instructions — false positivePE3Reaches for credential files — false positive
How this review was made
Sonnet 5 read the files above on 7 September 2026 and answered three questions: is it dangerous to whoever installs it, is each scanner finding real, and what should the installer know. The verdict is bound to the file's hash; when the file changes, it is scanned afresh and reviewed again. A script that changes while the definition does not is not re-reviewed — that is a known gap. How the scan and the review work.