Closes five gaps in the ultrawork prompt versus codex-plugins' parallel
directive, applied to all three model variants (default/Claude, GPT, Gemini)
with prompt-engineering entropy gate (each addition encodes a distinct
binding boolean, not narrative reinforcement):
1. TDD-MANDATORY (was conditional "when test infrastructure exists"):
every production change follows RED -> GREEN -> SURFACE. Failing test
first, capture assertion msg, smallest change to flip green, exercise
real surface, capture artifact. Exemption whitelist: formatting /
comment-only / version bump / rename-only, each must be justified
in writing; unjustified exemption = rejection.
2. Scenario contract (was free-form Test Plan Template): require 3+
scenarios upfront covering happy path, edge (boundary / empty /
malformed / concurrent), adjacent-surface regression. Each scenario
binds a binary pass condition, a real-surface artifact source, and
a test file + test id written test-first.
3. RED->GREEN evidence capture (was "all tests pass"): every scenario
requires TWO captured artifacts -- RED assertion msg before the
change AND GREEN assertion msg after -- alongside the real-surface
artifact (tmux / curl / browser / Playwright / computer-use /
CLI stdout / parsed config / DB diff). Tests are the floor (always
required); surface artifact is the ceiling (also required).
4. Durable notepad: mktemp -t ulw-*.md with append-only sections
(Plan, Scenarios, Now, Todo, Findings, Learnings). Survives context
loss; resume by re-reading.
5. Reviewer gate: trigger when user said strictly / rigorously /
"deeply", or task touches 3+ files / 20+ turns / 30+ min, or it is
refactor / migration / perf / security work. Reviewer verdict is
binding ("looks good but..." = rejection). Loop until unconditional
approval.
Plus: TODO format upgraded from vague "track every step" to atomic
`path: <action> for <scenario-id> -- verify by <check>` with a GOOD
test-first / impl pair example and a BAD list including
"production code before its failing test".
Per-variant adaptation:
- default.ts (Claude): full structured sections.
- gpt.ts (GPT-5.x): outcome-first prose, shorter prose per gpt-5.5 guide.
- gemini.ts: explicit enforcement framing + anti-optimism checkpoint
upgraded with a TDD-violation question (#7).
Verified by:
- bun test src/hooks/keyword-detector/ (119 pass / 0 fail).
- lsp_diagnostics clean on all three files.
- Module-load smoke test confirms each exported message string parses
and contains the new section anchors (TDD MANDATORY, SCENARIO
CONTRACT, DURABLE NOTEPAD, REVIEWER GATE).
Char deltas (directive body only):
- default 13646 -> 17144 (+26%)
- gpt 6740 -> 9215 (+37%, was the leanest start)
- gemini 14196 -> 16136 (+14%)
Existing tests only assert presence of "ULTRAWORK MODE ENABLED!" which
is preserved verbatim in every variant.
The lead currently leaves teams alive after the task list drains because
none of the prompt surfaces tell it WHEN to close or HOW. omx-style
'self-closing' behavior was missing for four reasons (diagnosed via
prompt-engineering A/B/C: wrong / misframed / missing):
1. builtin team-mode skill 'Lifecycle' (B+C): 'phase ends / shape
outgrown' is qualitative, so the model maps it to 'wait for user'.
Step 6 jumped to team_delete without the request/approve pair the
tool contract requires. Replaced with a 'Closure Contract' (a
computable predicate over team_task_list + team_status) and an
explicit 'Closure Sequence' (request -> approve -> delete, with
force=true reserved for unrecoverable paths only).
2. TEAM_MESSAGE keyword injection (C): spent 100%% of its one-shot
budget on routing ('do not substitute delegate_task'), 0%% on
closure. Added the same closure rule in compressed form. Kept the
'NEVER substitute with delegate_task' literal that
keyword-detector/index.test.ts depends on.
3. team-mode-status-injector body (C): the only per-session injection
for team mode had no closure obligation. Replaced the optional
'load the team-mode skill ... otherwise use the team_* tools'
sentence with a 'Closure invariant' clause that ties the check to
every team_task_update.
4. member-guidance Wrap-up (A+B): step 3 said 'so the lead can decide
whether to request shutdown', but team_shutdown_request is
lead-only - members cannot initiate it. Step ordering also placed
the completion message before team_task_update, so the lead's
closable check would see stale data. Reordered to
task_update -> check task_list for new work -> if nothing left,
send a single 'closure-ready' message and idle. Test assertion
updated to match the new accurate contract.
Also: stripped Korean alternation from TEAM_PATTERN per directive
('절대로 코드 내에 한국어 적지 마라'). Pattern is now
/\\bteam[\\s_-]?mode\\b/i. Removed 4 Korean test cases
(2 positive triggers + 2 false-positive guards) that the pattern no
longer needs to defend, and updated the keyword-detector AGENTS.md
row.
Net: -71 lines across prompt surfaces. The Closure Contract is the
only addition; everything else tightened.
Tests: 428/428 pass across src/features/team-mode/,
src/features/builtin-skills/, src/hooks/keyword-detector/,
src/hooks/team-mode-status-injector/, src/hooks/team-mailbox-injector/,
src/hooks/team-tool-gating/, src/hooks/team-session-events/.
LSP: no errors introduced (one pre-existing error in
keyword-detector/index.test.ts confirmed pre-existing on dev).
The hyperplan trigger \b(hyperplan|hpp)\b/i matched 'hpp' inside common C++ header references like 'check interface.hpp' or 'open buffer.hpp'. The leading '.' is a non-word character, so \b is already satisfied and the false positive fires the hyperplan-mode prompt on routine code questions.
Split the alternation so 'hpp' additionally requires that the preceding character is neither a word character nor a '.'. This preserves every existing trigger ('hpp do this', '/hpp ...', mid-sentence usage, mixed case) while rejecting filename uses of the .hpp extension. The longer 'hyperplan' keyword keeps the original \b boundary semantics.
Reproduction (added regression tests):
- 'please help to check interface.hpp' must NOT fire
- 'open src/include/audio/buffer.hpp and fix the leak' must NOT fire
All 14 cases in hyperplan.test.ts pass (12 existing + 2 new), broader keyword-detector suite stays green (92 pass), typecheck clean.
When default_mode.ultrawork is enabled, inject the ultrawork behavioral
instructions into the system prompt (experimental.chat.system.transform)
instead of prepending them to the user's visible chat message.
- system-transform.ts: No-op handler now injects ultrawork message into
output.system[] when defaultMode.ultrawork is active. Checks for existing
<ultrawork-mode> tag to avoid re-injection after compaction.
- keyword-detector/hook.ts: Removes the visible text injection
(output.parts[].text prepend) for default_mode.ultrawork path. Only
shows the toast as the user-facing indicator.
- plugin-interface.ts: Wires getUltraworkMessage from keyword-detector
into createSystemTransformHandler.
Result: Ultrawork mode activates silently via system prompt — the chat
stays clean, and only a toast shows 'Ultrawork Mode Active'.
Add new `default_mode` config section with two boolean fields:
- `ultrawork`: Auto-inject ultrawork mode prompt on main session start
without requiring the "ultrawork"/"ulw" keyword. Wired through the
keyword-detector hook — injects once per session, respects existing
guards (non-OMO agents, planner agents, subagent sessions).
- `ralph_loop`: Auto-start ralph loop on first main session message
without requiring /ralph-loop or /ulw-loop commands. When ultrawork
is also enabled, the loop starts in ultrawork mode.
Usage:
```jsonc
{
"default_mode": {
"ultrawork": true, // Always get ultrawork prompt on start
"ralph_loop": true // Auto-start ralph loop
}
}
```
Files: 7 modified/added, ~65 LOC added.
Add optional enabled_expansions field to keyword_detector config schema.
When set, acts as an allowlist - only those expansion types fire.
Empty array disables all expansions. Absent field keeps all enabled (backward-compatible).
Also supports coexistence with disabled_keywords denylist for fine-grained control.
- src/config/schema/keyword-detector.ts: add enabled_expansions field
- src/hooks/keyword-detector/detector.ts: apply allowlist filter in detectKeywordsWithType
- src/hooks/keyword-detector/hook.ts: pass enabled_expansions from config
- src/hooks/keyword-detector/index.test.ts: add 4 tests for enabled_expansions behavior
- assets/oh-my-opencode.schema.json: regenerate schema
When user types /hyperplan something, two pipelines race for the message:
1. keyword-detector hook (chat.message): the regex \b(hyperplan|hpp)\b
matches /hyperplan because \b is satisfied by the / boundary, so the
hook prepends <hyperplan-mode> to the text part.
2. auto-slash-command hook (chat.message, runs immediately after):
detectSlashCommand() checks `text.trimStart().startsWith("/")`. After
keyword-detector's prepend, the part now starts with <hyperplan-mode>
and the slash check fails, so the builtin command template
(with $ARGUMENTS substituted) is never injected.
The visible symptom: /hyperplan refactor X never runs the actual
HYPERPLAN_TEMPLATE - the user sees only the keyword-detector wrapper, which
is similar but not identical, and the slash command's $ARGUMENTS payload
is silently lost.
Fix: at the top of the keyword-detector hook, after isSystemDirective() but
before any keyword scan, bail out if the prompt text starts with a slash
command (^\s*\/[a-zA-Z][\w-]*\b). Slash commands are explicit invocations
and own their own mode-injection path; the keyword detector must not race
them. Free-form mentions like "hyperplan: refactor X" still trigger
keyword detection - only the leading-slash form is suppressed.
Lock the contract with three regression tests:
- /hyperplan refactor ... must NOT inject <hyperplan-mode>
- /hpp investigate ... must NOT inject (shorthand slash command)
- "hyperplan: refactor src/auth/handler.ts" still injects (free-form)
Generated with assistance of [OhMyOpenCode](https://github.com/code-yeongyu/oh-my-opencode).
Adjacent "hpp ulw" or "ulw hpp" (8 form combinations: short/long, both
orders) triggers a fused mode that suppresses the standalone ultrawork
and hyperplan banners and toasts in favor of one combo banner. The
combo banner explicitly preserves hyperplan's mandatory adversarial
workflow contract (do NOT improvise, do NOT skip rounds) instead of
silently downgrading it.
Suppression runs as a named pipeline step (suppressComboStandalones)
immediately after detection and before all consumers (planner filter,
session filters, toasts, message injection), so standalone toast checks
naturally see the already-suppressed list. Combo is allowed in non-main
sessions like ultrawork, filtered for planner agents like both
standalones, and blocked in subagent sessions via the existing gate.
disabled_keywords uses the intersection rule: disabling either
"ultrawork" or "hyperplan" also disables the combo, so no flavored
content leaks via the combo embedding when either base keyword is
disabled.
Includes a same-PR refactor of KEYWORD_DETECTORS from {pattern, message}
to {type, pattern, message} tuple shape, dropping the parallel hardcoded
types array in detector.ts that previously coupled type assignment to
registry index. Future detector additions can no longer silently corrupt
DetectedKeyword.type via reorder or insertion.
10 behavioral contract tests in hyperplan-ultrawork.test.ts cover both
trigger orders, non-adjacent rejection, suppression of injection and
toast, intersection-rule disable behavior, session/agent policy, and
ultrawork variant routing through the combo. The pre-existing combined
"ultrawork hyperplan" assertion in hyperplan.test.ts is removed in
favor of the new file.
Plan distilled from a hyperplan adversarial review (5 members,
3 rounds: skeptic, validator, researcher, architect, creative).
Update both injection paths to reflect the 7-phase workflow with the
mandatory plan agent handoff in step 7. The keyword detector now lists
all 8 enumerated steps including the explicit "do not write the plan
yourself" guidance, and the /hyperplan command template references the
7-phase workflow count.
Generated with assistance of [OhMyOpenCode](https://github.com/code-yeongyu/oh-my-opencode).
Adds keyword_detector.disabled_keywords config so users can opt out of
specific keyword detectors individually without disabling the entire
keyword-detector hook. Allowed values: 'ultrawork', 'search', 'analyze',
'team'. Default empty/missing -> all four detectors active (no behavior
change for existing configs).
Motivation: an audit revealed search and analyze patterns trigger on
~30-60% of normal conversational user messages (e.g. 'how to', 'why is',
'show me', '왜', '어떻게'). The disable list is the immediate kill switch
while the patterns themselves are tightened in a separate PR.
Schema follows the existing per-feature config block convention shared
by team_mode, ralph_loop, runtime_fallback, and comment_checker. The
KeywordType enum (z.enum) lives next to the config schema and is
re-imported by the detector to keep the union type in lockstep with the
schema.
Threading:
pluginConfig.keyword_detector
-> create-transform-hooks.ts (factory wiring)
-> createKeywordDetectorHook(config)
-> detectKeywordsWithType(text, agent, model, disabledKeywords)
-> Set-based filter at the source-of-truth detector
Adds 8 regression tests covering per-keyword disable, multi-keyword
disable, partial disable (one keyword off, another still firing),
ultrawork toast suppression, undefined config, and empty array.
Detects user invocations of team-mode work across English and Korean
('team mode', '팀 모드', '팀으로') and injects a concise English directive
instructing the LLM to orchestrate via team_* tools (team_create ->
team_task_create + team_send_message), forbidding delegate_task
substitution and fallbacks.
The Korean variants use a Hangul-syllable negative lookbehind (가-힣) so
that '스팀으로 게임 켜줘' does not falsely match '팀으로' and '스팀모드' does
not falsely match '팀모드'.
Follows the existing folder pattern (mode/default.ts + mode/index.ts)
shared by ultrawork/, search/, and analyze/. The hook orchestration in
hook.ts handles the new keyword type generically through the shared
KEYWORD_DETECTORS array, so existing guards (non-OMO agent skip,
non-main session filter, system-reminder strip, code-block strip) all
apply automatically.
Adds 7 regression tests covering English/Korean trigger forms, the
Hangul-prefix false-positive guard, the bare-'team' negative case, and
non-main-session filtering.
- #3354: Coerce data.name to String in loadSkillFromPath/loadSkillFromPathAsync
to prevent crash when YAML parses numeric skill names (e.g., name: 12306)
- #3416: Add required run_in_background parameter to all task() examples in
ultrawork prompts (default, gpt, gemini, planner) to match tool schema
- #3379/#3417/#3418/#3337/#3335: Strip ZWSP (U+200B) before agent name
comparisons in agent-tool-restrictions, sync-prompt-sender, tool-execute-after,
tool-execute-before, oracle-verification-detector, call-omo-agent,
recovery-prompt-config, and agent-variant to prevent ZWSP-prefixed display
names from breaking exact-match lookups
YeonGyu-Kim