인터페이스는 현지화되어 있으며, Skill 소개와 설명 문서는 선택한 언어로 AI가 자동 번역합니다. 코드·명령어·경로·전문 용어는 원문을 유지하며 원작자 문서가 우선합니다.
Skills 둘러보기
문서 및 사무 · SKILL.md

platform-actions

platform-actions

Document OASIS platform actions, platform settings, recommendation behavior, SQLite schema, trace behavior, group/interview/report/product workflows, and DB diagnostics.

Independent GitHub snapshot · Automated classification · Original documentation governs

Skill 문서

GitHub

name: platform-actions description: "Document OASIS platform actions, platform settings, recommendation behavior, SQLite schema, trace behavior, group/interview/report/product workflows, and DB diagnostics." disable-model-invocation: true metadata: disco-role: operating license: Apache 2.0

Platform Actions

Use this sub-skill when you need to choose the correct ManualAction for OASIS platform behavior, understand the Platform constructor and default presets, read or summarize the SQLite database, or debug why a platform action did not land in the expected table or trace row.

Route elsewhere

  • Agent or profile creation flows -> agent-profiles
  • Full simulation lifecycle, step orchestration, and run control -> simulation-workflows
  • Visualization and legacy result analysis -> experiments-analysis

What this sub-skill covers

  • ActionType values and the built-in Twitter/Reddit default action sets
  • Exact ManualAction.action_args keys for platform, comment, group, report, product, search, trend, refresh, and interview actions
  • Platform.__init__ configuration, default platform presets, Channel, Clock, and RecsysType
  • SQLite tables, trace payloads, and recommendation-table behavior
  • Action-specific failure modes and DB inspection with scripts/oasis_db_summary.py

Operating flow

  1. Start in references/action-reference.md to pick the action type and its argument names.
  2. Use references/platform-and-recsys.md to confirm the platform preset, recsys type, channel, clock, and self-rating settings.
  3. Use references/database-and-traces.md to map the action to the SQLite tables and trace payload.
  4. If the action failed or returned an empty result, use references/troubleshooting.md.
  5. Inspect a live DB with scripts/oasis_db_summary.py before guessing at schema or trace state.

Notes

  • INTERVIEW is usually a manual or harness-driven action; keep it out of LLM available_actions unless you intentionally want the model to choose it.
  • SIGNUP, UPDATE_REC_TABLE, and EXIT are infrastructure-level actions. This sub-skill only routes them; it does not cover agent/profile creation or full environment lifecycle.
  • When a post id belongs to a repost or quote, several actions resolve to the root post internally. See the action reference before debugging duplicate or missing-post errors.
저장한 Skills