Menu
Update progress and next-step docs so the next agent or session can resume with minimal context loss.
Install with Codex or Claude Copy this prompt, paste it into Codex, Claude, or another assistant, and let it review the skill page and install it for you.
Based on SOC occupation classification
Run one bounded episteme Event end-to-end โ branch, implement, verify, PR, merge, sync, doctor โ with the repo's release and hook machinery handled correctly. Use when starting or closing a unit of work in the episteme repo (or any repo adopting its flow). Triggers on "new event", "์ด๋ฒคํธ ์์", "PR๊น์ง", "mergeํ๊ณ sync", "release", "๋ฆด๋ฆฌ์ฆ", "๋ธ๋์น ํ์ ์์ ".
Author the episteme gate artifact (.episteme/reasoning-surface.json) correctly on FIRST contact โ schema, lens order, honest values, strict-vs-advisory, and when to escalate to epistemic-interrogation instead. Use the moment a PreToolUse hook reports a missing or stale Reasoning Surface, or BEFORE starting a session's first high-impact operation. Triggers on "reasoning surface", "Episteme Strict Mode", "surface stale", "blocked by episteme", "๊ฒ์ดํธ ๋งํ", "์ํผ์ค ๊ฐฑ์ ".
Devil's Advocate stress-testing for code, architecture, PRs, and decisions. Surfaces hidden flaws through structured adversarial analysis with metacognitive depth. Use for high-stakes review, stress-testing choices, or when the user wants problems found deliberately. NOT for routine code review (use /code-review). Triggers on "์คํธ๋ ์ค ํ ์คํธ", "stress test", "devil's advocate", "๋ฐ๋ก ", "์ด๊ฑฐ ๊ด์ฐฎ์", "๋ฌธ์ ์์๊น", "๊น์ ๋ฆฌ๋ทฐ", "critical review", "adversarial".
Senior-researcher interrogation of a load-bearing decision or conclusion โ decompose into claims, verify the load-bearing ones in a fresh context with external evidence, argue the opposition, name the weakest link, pre-commit disconfirmation, record the verdict artifact the episteme gate reads.
Design and run unattended or long-horizon loops (overnight hardening, drain queues, iterative sweeps) with a written invariant the harness re-reads every iteration โ no manual re-pasting, trustworthy metrics, explicit stop conditions, and correct model routing. Replaces bounded-loop-runner. Use for "overnight", "unattended", "keep iterating until", "๋ฐค์ ๋๋ ค", "๋ฃจํ ๋๋ ค", ralph-loop or /loop driven work.
Prevents premature execution on ambiguous requests. Analyzes request clarity using 5W1H decomposition, surfaces hidden assumptions, and generates structured clarifying questions before work begins. Use at the start of any non-trivial task, or when a request could be interpreted multiple ways. Triggers on "๋ญ ์ํ๋๊ฑด์ง", "์๊ตฌ์ฌํญ ์ ๋ฆฌ", "clarify", "what exactly", "scope", "requirements", "์ ํํ ๋ญ", "before we start".
| name | progress-handoff |
| description | Update progress and next-step docs so the next agent or session can resume with minimal context loss. |
Use this skill near the end of a work session or after substantial progress.
Convention check first: if the project's AGENTS.md (or operating contract)
declares its own handoff convention โ different file names, a different form โ
follow the project's declaration. Repo files own repo mechanics; the two
moves below are the episteme default for projects that declare none.
Handoff is two moves, and only two:
REPLACE docs/NEXT_STEPS.md. Overwrite it with the current state โ do not
append. Include:
Append exactly ONE line to docs/EVENTS.md โ one row in the history
index (| E<N> | date | one-line what happened | refs |). Nothing more; no
narrative body. Verbatim detail, if it must be kept, goes in archive/.
docs/PROGRESS.md is retired in this pattern. Do not create it and do not
write to it โ its append-log shape was the disease compaction cured. The
one-line-per-event index (EVENTS.md) plus the REPLACE-form NEXT_STEPS.md
carry everything it used to.
Favor concise operational handoffs over narrative summaries.
