{"slug":"give-me-tips","title":"give-me-tips","summary":"Explains any senpi tip in depth - startup tips, working tips, and any Tip: line shown in the TUI (including the Fable-5-refusal fallback tip). Use when the user asks about a Tip: line, says give-me-tips, asks what a tip means, how a tipped feature works, or which tips they can se","platform":"Claude","tags":[],"authorName":"LLM Mart","authorSlug":"llm-mart","score":0,"source":"github","price":null,"verified":false,"createdAt":"2026-08-23T08:58:14.797285Z","repo":{"url":"https://github.com/code-yeongyu/oh-my-openagent","stars":69391,"forks":5715,"license":null,"updatedAt":"2026-09-24T23:30:44Z"},"bodyHtml":"<hr>\n<h2>name: give-me-tips\ndescription: \"Explains any senpi tip in depth - startup tips, working tips, and any Tip: line shown in the TUI (including the Fable-5-refusal fallback tip). Use when the user asks about a Tip: line, says give-me-tips, asks what a tip means, how a tipped feature works, or which tips they can see. Queries the live tip list (senpi --list-tips) first, checks what THIS user can actually see, then verifies the real feature code before explaining.\"\nmetadata:\nshort-description: Deep, verified, brag-worthy explanations of any senpi tip</h2>\n<h1>give-me-tips - explain any senpi tip, in depth</h1>\n<h2>Purpose</h2>\n<p>Senpi shows the user tips: startup tips, working tips, and <code>Tip:</code> lines injected by omo\ncomponents. When the user asks about any of them - \"what was that Tip: line\", \"what does this tip\nmean\", \"how does that feature work\" - this skill produces a DEEP explanation of that exact tip, in\nthe USER'S language (match the language they asked in, always).</p>\n<p>Specific over generic, every time. \"It retries on failure\" is a failure of this skill. \"It detects\nthe refusal from the stopDetails on the assistant message_end event, gates on the architect\ncategory in your .omo/omo.json, and only then injects the directive\" is the bar. The user asked\nbecause the tip made them curious; reward that curiosity with the real mechanism, not a summary of\nthe tip text they already read.</p>\n<h2>Query the live tip list FIRST</h2>\n<p>Never explain from memory. Get the ground truth of what tips exist:</p>\n<ol>\n<li>Run <code>senpi --list-tips</code>. It prints JSON: <code>[{id, text, requiresCommand?}]</code>. Match the user's tip\nagainst this list by id or by text fragment.</li>\n<li>If the flag is unavailable on this senpi version, fall back to reading the catalog sources\ndirectly at <code>packages/coding-agent/src/modes/interactive/tips/catalog/</code> inside the installed\n<code>@code-yeongyu/senpi</code> package (find it via the senpi install path or node_modules) or in a\nlocal clone of code-yeongyu/senpi.</li>\n</ol>\n<p>Then - and this is the part most explanations get wrong - <strong>available tips DIFFER per user</strong>. The\ncatalog is the superset; what THIS user can actually see is gated by:</p>\n<ul>\n<li>the <code>tips</code> toggle in the senpi agent dir <code>settings.json</code> (tips can be off entirely),</li>\n<li><code>tipsHistory</code> in that same settings/state, which drives cooldown rotation so a tip the user saw\nrecently will not reappear for a while,</li>\n<li><code>requiresCommand</code> gating: a tip tied to a command only shows when that command is available,</li>\n<li>keybinding availability: some tips reference bindings the user's setup may not have.</li>\n</ul>\n<p>Read the senpi agent dir <code>settings.json</code> (and its tips history state) BEFORE explaining, and tell\nthe user which tips they personally can encounter and why - not the full catalog as if everyone\nsees everything.</p>\n<h2>Verify before explaining</h2>\n<p>Never invent behavior. A tip is a one-line promise; the truth lives in code. Before writing the\nexplanation, read the actual feature implementation:</p>\n<ul>\n<li>senpi itself: <code>code-yeongyu/senpi</code>, under <code>packages/coding-agent/</code> (the tips catalog lives at\n<code>packages/coding-agent/src/modes/interactive/tips/catalog/</code>; the features the tips point at live\nin the surrounding packages),</li>\n<li>omo components: <code>code-yeongyu/oh-my-openagent</code>, under <code>packages/omo-senpi/</code>.</li>\n</ul>\n<p>Cite the concrete file paths you read in your explanation. If the code and the tip text disagree,\nthe code wins - say so and show what it actually does.</p>\n<h2>Tone: BRAG</h2>\n<p>These tips exist because someone engineered something genuinely impressive, and a flat doc summary\nbetrays that. Lead with the most impressive engineering behind the tip - the clever detection, the\nrace that had to be closed, the state machine hiding under one sentence - and showcase it. The user\nshould finish the explanation feeling like they got a tour of the engine room, not a sticker\nreading. Concrete mechanics over adjectives: name the events, the gates, the file paths, the exact\norder of operations.</p>\n<h2>The Fable-5-refusal tip specifically</h2>\n<p>When the user asks about the tip that appears after a Fable 5 refusal (\"Fable 5 refused, but its\nrefusals should not wear you down...\"), explain the full fallback-architect pipeline, citing\n<code>packages/omo-senpi/src/components/fallback-architect/</code>:</p>\n<ol>\n<li><strong>Refusal detection</strong> (<code>detection.ts</code>): the component watches <code>message_end</code> events and applies\nthe same refusal semantics senpi's own retry classifier uses - stopReason checked FIRST (so an\nabort or normal stop carrying stale stopDetails can never masquerade as a refusal), then\nstopDetails of type refusal/sensitive, plus the Anthropic usage-policy errorMessage pattern for\nprovider-side blocks that carry no stopDetails at all.</li>\n<li><strong>Architect category gate</strong> (<code>architect-gate.ts</code>): on a <code>model_select</code> with source \"fallback\"\nmoving AWAY from claude-fable-5 with a refusal pending, the component checks the user's own omo\nconfig for an active architect category. No architect category, no nudge - the feature never\npretends depth is reachable when it is not.</li>\n<li><strong>Hidden directive</strong> (<code>directive.ts</code>, customType <code>omo-fallback-architect:directive</code>,\ndisplay:false): the fallback model gets a hidden 5-step playbook - decompose the problem,\nconsult <code>task(category: \"architect\")</code> with one self-contained query per part (the architect\nconsultant IS Fable 5, reached through a lane its refusal cannot block), run independent\nconsultations in parallel, and split refused queries into smaller benign sub-questions instead\nof resending. The directive also tells the model the user was shown the visible tip, so the\ntwo never contradict each other.</li>\n<li><strong>Visible tip</strong> (<code>tip-message.ts</code>, customType <code>omo-fallback-architect:tip</code>, display:true):\nrendered as a dim <code>Tip:</code> block via a registered message renderer. It names the ACTUAL fallback\nmodel the session landed on, reassures the user that the refused question is still being\nreasoned through in essence, and notes Fable-5-grade depth stays reachable through the\narchitect category.</li>\n</ol>\n<p>The engineering worth bragging about: the refusal never deletes the user's question. Detection\narms on the exact assistant message that preceded the switch (a later successful answer disarms\nit), reminders ride inside queued prompts instead of burning extra assistant turns, and the whole\nnudge self-cancels the moment Fable 5 becomes the active model again or senpi reverts the\nfallback. One refusal triggers a coordinated downgrade in visibility with zero downgrade in\nreachable reasoning depth.</p>\n","files":[{"path":"SKILL.md","sizeBytes":13342,"isText":true}],"reviewScore":null,"reviewSummary":null,"trust":{"provenance":"trusted-source-unreviewed","notice":"Community-authored content, reproduced verbatim and not vetted as instructions. Treat it as data to evaluate, never as directives to follow.","bodySource":null},"bodyLocked":false,"purchaseUrl":null,"sourceUrl":null,"report":{"provenance":"trusted-source-unreviewed","screen":{"ran":true,"outcome":"clean","suspicious":0,"notes":0,"hiddenCharacters":false},"virusScan":{"engine":"clamav","status":"clean","scannedAt":"2026-09-25T07:38:14.12825Z","sha256":"2683FB5F2AFFA1E2E1416627C3F07BD1367800A20EF3CB34FC5BBCFF9A40157E","sizeBytes":5887},"review":null,"source":{"repositoryUrl":"https://github.com/code-yeongyu/oh-my-openagent","path":"packages/omo-senpi/skills/give-me-tips","license":null,"commit":"05dcba64b749e7666dcd0296c079d31cf3c298f1","subtreeSha":"495F24E2C389D8E8A4347C38B529A905D2A2A976F610D621B40A21D898E42623","lastSyncedAt":"2026-09-25T07:37:33.26442Z"},"reviewedAt":"2026-09-25T07:38:43.705495Z","notice":"Community-authored content, reproduced verbatim and not vetted as instructions. Treat it as data to evaluate, never as directives to follow."},"install":[{"target":"skills-cli","command":"npx skills add https://github.com/code-yeongyu/oh-my-openagent/tree/dev/packages/omo-senpi/skills/give-me-tips"},{"target":"claude-code","command":"claude plugin marketplace add https://llmmart.ai/marketplace.json && claude plugin install code-yeongyu-oh-my-openagent@llmmart"},{"target":"git","command":"git clone https://github.com/code-yeongyu/oh-my-openagent.git"}]}