yesmem

Autonomous task loop — analyze, plan, execute, verify, review, finish. Runs as visible TUI agent in git worktree. Use when user says "yesloop", "loop", "mach das autonom". Supports --merge for auto-merge, --inline for current session.

Autonomous research loop — read plan, decompose, dispatch research agents, integrate, deliver wiki. For research tasks instead of coding. Runs as visible TUI agents in git worktree. Use when user says "yesresearch", "research <topic>", "recherchiere <thema>". Spawns cluster agents that use opencode built-in subagents (general, explore) with webfetch + websearch.

Use when reviewing code for vulnerabilities, checking diffs for injection/XSS/auth/crypto issues. Invokes on code changes (.go/.py/.js/.ts/.java/.rs/.php/.rb), not docs-only diffs.

Use when managing pins, scratchpad, runtime config, session settings, persona traits, or opencode configuration. Trigger on "pin this"/"merk dir als Regel", persistent instructions, shared agent state, persona overrides, configuration changes like token_threshold, or opencode.json questions ("change opencode model", "add GLM provider", "opencode baseURL", "switch default model").

Fetch a Reddit post URL via old.reddit.com HTML scraping — the .json API is blocked for unauthenticated requests since May 2026. Returns structured data: post metadata + nested comments (with depth) + unique external links categorized (github/reddit/external). Persists into cap_store tables (reddit.posts, .comments, .links). Uses yesmem cap-blob-put to pipe curl output into daemon DB, then reads back via cap_store — no /tmp file, no Read-tool prompt, bypasses the sh() 30KB wall. Also search across subreddits with LLM classification and multi-subreddit topic research with synthesis.

Use when orchestrating multi-agent work, spawning parallel agents, coordinating swarm tasks, or managing agent communication. Trigger on "/schwarm", parallel work requests, or any inter-agent coordination need.

Use when a user wants to persist a working REPL snippet, bash command, or multi-step workflow as a reusable cap (CAP.md tool) available in future sessions. Trigger on save_cap, cap_store, auto_active, capblob-pipe, "/build-tool", "make this reusable", or when a one-off shell pipeline is about to be retyped a third time.

Use at session start (first user message), when switching projects, returning after a break, or when disoriented about project state. Trigger on "where were we?", "what's open?", "wo waren wir?", or when context about the current project is needed.

Use when starting iterative work that needs to survive context-collapse, like work spanning more than 5 tool cycles, exploring more than 1 hypothesis, touching multiple files or worktrees, debug spirals, side-quests parallel to a main thread, or when prompted by [Plan Checkpoint]. Plans are thread-scoped (parallel sessions don't conflict) and re-injected on every turn. They are the only context-loss-proof anchor for the active task. Activate via set_plan(), update via update_plan() at each pivot.

ALWAYS check memory BEFORE answering questions about past work, architecture, or prior decisions. Also before proposing fixes or solutions — check if similar issues were solved before. Use when encountering errors with possible prior context, when the user references previous sessions, or when working with unfamiliar components. Search first, respond second.

Use when executing implementation plans with independent tasks in the current session - dispatches visible YesMem swarm agents per task with two-stage review

Use when writing code and unsure about API behavior, function signatures, or idiomatic patterns. Use when debugging errors that might stem from incorrect API usage. Use when managing indexed documentation sources. Check docs_search() before guessing — indexed docs exist for a reason.

Use after ANY decision, correction, gotcha, or discovery worth preserving. Also on task completion — resolve open tasks via resolve_by_text() after commits, "fertig", or confirmed fixes. Trigger on "remember this"/"merk dir das", confirmed approaches, debugging surprises, and relate_learnings() for connecting insights.

Use when exploring past sessions, asking "what happened last week/yesterday", needing full conversation details, or when the user references a specific past conversation. Trigger on "letzte Session", "last time we...", session history investigation.