hooks/pre-task-ritual.sh has no test at all. It is one of only two bash-shebang
hooks in this repo; the other, hooks/secret-guard.sh, was just found to have been dead in
CI for as long as its suite was silently skipped (see skills 4edf42c).
Why this specific hook matters: it enforces the Rule 0 pre-task ritual — brain query, skill
load, failing test first. Its failure mode, like any hook that cannot start, is to produce
no output, and a hook that produces no output does not block anything. So "enforcing" and
"dead on line 1" are indistinguishable from the outside, and there is currently nothing
that would tell them apart.
invoke it as a program, never sh "$HOOK" — that override is exactly what hid the
secret-guard breakage, since it substitutes a different interpreter than production uses
assert the exec bit, since direct invocation depends on it
one allow case and one block case, so a hook that never blocks fails the suite
Not urgent — no evidence it is currently broken — but it is untested by inspection, and the
sibling hook's failure showed that "untested and probably fine" and "dead" look the same
from here.
`hooks/pre-task-ritual.sh` has **no test at all**. It is one of only two bash-shebang
hooks in this repo; the other, `hooks/secret-guard.sh`, was just found to have been dead in
CI for as long as its suite was silently skipped (see skills `4edf42c`).
Why this specific hook matters: it enforces the Rule 0 pre-task ritual — brain query, skill
load, failing test first. Its failure mode, like any hook that cannot start, is to produce
no output, and a hook that produces no output does not block anything. So "enforcing" and
"dead on line 1" are indistinguishable from the outside, and there is currently nothing
that would tell them apart.
Minimum worth having, mirroring `tests/secret_guard_test.sh`:
- invoke it **as a program**, never `sh "$HOOK"` — that override is exactly what hid the
secret-guard breakage, since it substitutes a different interpreter than production uses
- assert the exec bit, since direct invocation depends on it
- one allow case and one block case, so a hook that never blocks fails the suite
Not urgent — no evidence it is currently broken — but it is untested by inspection, and the
sibling hook's failure showed that "untested and probably fine" and "dead" look the same
from here.
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
hooks/pre-task-ritual.shhas no test at all. It is one of only two bash-shebanghooks in this repo; the other,
hooks/secret-guard.sh, was just found to have been dead inCI for as long as its suite was silently skipped (see skills
4edf42c).Why this specific hook matters: it enforces the Rule 0 pre-task ritual — brain query, skill
load, failing test first. Its failure mode, like any hook that cannot start, is to produce
no output, and a hook that produces no output does not block anything. So "enforcing" and
"dead on line 1" are indistinguishable from the outside, and there is currently nothing
that would tell them apart.
Minimum worth having, mirroring
tests/secret_guard_test.sh:sh "$HOOK"— that override is exactly what hid thesecret-guard breakage, since it substitutes a different interpreter than production uses
Not urgent — no evidence it is currently broken — but it is untested by inspection, and the
sibling hook's failure showed that "untested and probably fine" and "dead" look the same
from here.