Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add tellang/triflux --skill tfx-autogit clone --depth 1 https://github.com/tellang/trifluxWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/tellang/triflux/tfx-auto)<a href="https://agentmods.dev/skills/tellang/triflux/tfx-auto"><img src="https://agentmods.dev/badge/skills/tellang/triflux/tfx-auto/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/tellang/triflux/tfx-auto"><img src="https://agentmods.dev/badge/skills/tellang/triflux/tfx-auto.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00139 | $0.16879 |
| Opus 5 | $0.00069 | $0.08440 |
| Sonnet 5 | $0.00028 | $0.03376 |
| Haiku 4.5 | $0.00014 | $0.01688 |
Grade C, and why
tfx-auto scanned grade C with 2 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 10d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Downloads and executes remote codehighSupply chain
curl | sh runs whatever the server returns today, which is not necessarily what it returned when this was reviewed.
- codex: `npm install -g @openai/codex` | antigravity: `curl -fsSL https://antigravity.google/cli/install.sh | bash` Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- codex: `npm install -g @openai/codex` | antigravity: `curl -fsSL https://antigravity.google/cli/install.sh | bash` How it starts
The opening of the file, as written. The whole thing — 1,114 lines — stays where its author put it; the contents beside it link to each section on GitHub.
tfx-auto — 통합 CLI 오케스트레이터
ARGUMENTS 처리: 이 스킬이
ARGUMENTS: <값>과 함께 호출되면, 해당 값을 사용자 입력으로 취급하여 워크플로우의 첫 단계 입력으로 사용한다. ARGUMENTS가 비어있거나 없으면 기존 절차대로 사용자에게 입력을 요청한다.
Step 0: 스마트 라우팅 (tfx-auto 진입 시 자동 실행)
preamble에서 routing-weights.json을 읽고, 사용자 입력을 분석하여 dispatch 결정.
SLUG=$(basename "$(git rev-parse --show-toplevel 2>/dev/null)" 2>/dev/null || echo "unknown")
WEIGHTS_FILE="$HOME/.gstack/projects/$SLUG/routing-weights.json"
USER_MODE=""
if [ -f "$WEIGHTS_FILE" ]; then
USER_MODE=$(node -e "
const w=JSON.parse(require('fs').readFileSync('$WEIGHTS_FILE','utf8'));
const m=w.weights?.mode_bias||{};
const top=Object.entries(m).sort((a,b)=>b[1]-a[1])[0];
if(top && top[1]>0.3) console.log(top[0]);
" 2>/dev/null)
fi
echo "USER_PREFERRED_MODE: ${USER_MODE:-none}"
판단 기준 (우선순위 순):
-
명시 플래그 (최우선, 추론 스킵): ARGUMENTS 에
--cli/--mode/--risk-tier/--shape/--cli-set/--parallel/--retry/--isolation/--remote플래그가 있으면 분류/추론을 건너뛰고 플래그 값대로 즉시 dispatch. 자세한 플래그 동작은 아래 "플래그 오버라이드" 섹션 참조.--parallel swarm→ tfx-swarm 엔진 위임 (PRD 필요)--parallel N→ tfx-multi 엔진 위임 (auto: 리드 tmux면 interactive pane, 없으면 in-process)--cli codex|antigravity→TFX_CLI_MODE설정 + 단일 실행--mode deep→-t/--thorough동일 동작 (pipeline init)--risk-tier low|medium|high→ risk-tier 기준으로 verification 강도와 mode 결정--mode ...명시 시--risk-tier는 무시 (mode 우선)--mode consensus --shape debate|panel→ prompt ensemble fold 경로--mode live→tfx-live엔진 위임 (분해/트리아지 스킵)--retry ralph→ stderr 경고 후 bounded 3회 degrade (Phase 2 미구현)
-
사용자 명시 키워드 (플래그 없을 때):
- "swarm", "PRD 돌려", "격리해서", "충돌 없이 돌려" →
--parallel swarm --mode consensus --isolation worktree - "병렬", "동시에", "multi", "협업"(단순 동시 작업 의미) →
--parallel N --mode deep— 여기서 미리 swarm으로 단정하지 않는다. 실제 코드 변경이 여러 파일에 걸치면 "멀티 태스크 라우팅"의 자동 분류가 swarm으로 승격시킨다 ("병렬" 한 단어만으로 곧장 worktree 격리+3-CLI 합의까지 가는 건 과함) - "팀"은 모호하다: 서로 메시지를 주고받는 영속 협업을 원하면 정답은 OMC 네이티브
/team이고(tfx-auto 관할 밖), 그냥 여러 작업을 동시에 처리해달라는 뜻이면 바로 위--parallel N행을 따른다. "팀"만 보고 바로--parallel N으로 단정하지 말고 어느 쪽인지 애매하면 되묻는다 - "꼼꼼히", "제대로", "deep" →
--mode deep - "끝까지", "멈추지마", "ralph" →
--retry ralph - "codex로", "antigravity로" →
--cli codex또는--cli antigravity - "원격으로", "다른 기기에서", "리모트로 돌려" →
--remote <host>(--parallel swarm자동 동반. host 미지정 시hosts.json목록에서 질의) - "논쟁시켜", "서로 반박하게 해", "계속 대화하면서 풀게", "왕복으로 주고받게" →
--mode live(tfx-live peer고정). "협업"/"팀"과 겹쳐 보여도 이 쪽은 "논쟁/반박/대화를 계속 이어간다"는 뉘앙스가 명시적으로 있을 때만 해당
- "swarm", "PRD 돌려", "격리해서", "충돌 없이 돌려" →
What ships with it
1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 10d ago First seen · 1,114 lines · 139 tokens per session scan C f2890c43e984
tfx-auto is a skill published in the GitHub repository tellang/triflux (7 stars, last pushed yesterday), licensed MIT. It adds 139 tokens to every session and 16,879 once invoked, about $0.0007 per session on Opus 5. A static security scan graded it C with 2 findings (downloads and executes remote code, makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other skills, from other repositories
lucia-chat
Set two agent sessions talking through the Luciazero Agent Bus and watch it live in a terminal: pick the pair, open the windows, read the transcript. Use for "ให้ codex กับ claude คุยกัน" or "watch the bus".
ready
Make an unfamiliar repository agent-ready with a verify command, smoke tests, guardrails, and project notes. Use for repository setup, agentic engineering, verify commands, hooks, or allowlists; skip when verification and scope are already clear.
handoff
Write a state capsule so the next session — or a different agent/harness — can resume unfinished work without re-deriving context. Use when a session is ending with work incomplete, when the user says "handoff", "pack up", "ส่งต่อ", "continue tomorrow", when switching between Claude Code and Codex mid-task, or when…
done
Run the closeout ritual before handing back non-trivial work; full verification, revert-probe honesty, independent review, and scope reporting. Use before declaring completion, opening a PR, wrapping up a change, or "ปิดงาน".
retro
Record durable lessons, null results, and footguns after hard work or debugging. Use when the user asks for a retro, dead ends need preserving, a task disproves an approach, or "จดบทเรียน". Keep repo knowledge separate from machine-local memory.
show
Visualize code structure, changes, and verification evidence in the smallest useful view. Use for connections, flows, diffs, file maps, Mermaid diagrams, evidence maps, or focused HTML; show facts and label unknowns.