# Send Planning With Files to your agent Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually. ## Fast path - Download the package from Yavira. - Extract it into a folder your agent can access. - Paste one of the prompts below and point your agent at the extracted folder. ## Suggested prompts ### New install ```text I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Then review README.md for any prerequisites, environment setup, or post-install checks. Tell me what you changed and call out any manual steps you could not complete. ``` ### Upgrade existing ```text I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Then review README.md for any prerequisites, environment setup, or post-install checks. Summarize what changed and any follow-up checks I should run. ``` ## Machine-readable fields ```json { "schemaVersion": "1.0", "item": { "slug": "planning-files", "name": "Planning With Files", "source": "tencent", "type": "skill", "category": "开发工具", "sourceUrl": "https://clawhub.ai/wpank/planning-files", "canonicalUrl": "https://clawhub.ai/wpank/planning-files", "targetPlatform": "OpenClaw" }, "install": { "downloadUrl": "/downloads/planning-files", "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=planning-files", "sourcePlatform": "tencent", "targetPlatform": "OpenClaw", "packageFormat": "ZIP package", "primaryDoc": "SKILL.md", "includedAssets": [ "README.md", "SKILL.md", "templates/progress.md", "templates/task_plan.md", "templates/findings.md", "scripts/init-session.sh" ], "downloadMode": "redirect", "sourceHealth": { "source": "tencent", "slug": "planning-files", "status": "healthy", "reason": "direct_download_ok", "recommendedAction": "download", "checkedAt": "2026-05-07T04:42:22.412Z", "expiresAt": "2026-05-14T04:42:22.412Z", "httpStatus": 200, "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=planning-files", "contentType": "application/zip", "probeMethod": "head", "details": { "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=planning-files", "contentDisposition": "attachment; filename=\"planning-files-1.0.0.zip\"", "redirectLocation": null, "bodySnippet": null, "slug": "planning-files" }, "scope": "item", "summary": "Item download looks usable.", "detail": "Yavira can redirect you to the upstream package for this item.", "primaryActionLabel": "Download for OpenClaw", "primaryActionHref": "/downloads/planning-files" }, "validation": { "installChecklist": [ "Use the Yavira download entry.", "Review SKILL.md after the package is downloaded.", "Confirm the extracted package contains the expected setup assets." ], "postInstallChecks": [ "Confirm the extracted package includes the expected docs or setup files.", "Validate the skill or prompts are available in your target agent workspace.", "Capture any manual follow-up steps the agent could not complete." ] } }, "links": { "detailUrl": "https://openagent3.xyz/skills/planning-files", "downloadUrl": "https://openagent3.xyz/downloads/planning-files", "agentUrl": "https://openagent3.xyz/skills/planning-files/agent", "manifestUrl": "https://openagent3.xyz/skills/planning-files/agent.json", "briefUrl": "https://openagent3.xyz/skills/planning-files/agent.md" } } ``` ## Documentation ### Planning with Files Use persistent markdown files as your "working memory on disk." Based on context engineering principles from Manus. ### OpenClaw / Moltbot / Clawbot npx clawhub@latest install planning-with-files ### WHAT This Skill Does Treats the filesystem as persistent memory to overcome context window limitations. Three files track your state: FilePurposeUpdate Frequencytask_plan.mdPhases, progress, decisionsAfter each phasefindings.mdResearch, discoveries, decisionsAfter ANY discoveryprogress.mdSession log, test results, errorsThroughout session ### WHEN to Use Use for: Multi-step tasks (3+ steps) Research tasks requiring web search Building/creating projects from scratch Tasks spanning >5 tool calls Anything requiring organization across multiple files Tasks where losing context would cause rework Skip for: Simple questions Single-file edits Quick lookups Tasks completable in 1-2 actions Keywords: complex task, multi-step, research, build project, create application, plan, organize ### The Core Pattern Context Window = RAM (volatile, limited) Filesystem = Disk (persistent, unlimited) → Anything important gets written to disk. ### Phase 1: Create Planning Files Before starting ANY complex task, create all three files in your project root: Create task_plan.md — Copy from templates/task_plan.md Create findings.md — Copy from templates/findings.md Create progress.md — Copy from templates/progress.md ### Phase 2: Execute with Discipline Follow these rules during execution: The 2-Action Rule: After every 2 view/browser/search operations, IMMEDIATELY save findings to text files. Visual/multimodal content doesn't persist — write it down before it's lost. Read Before Decide: Before major decisions, read your plan file. This keeps goals in your attention window after many tool calls. Update After Act: After completing any phase: Mark phase status: in_progress → complete Log any errors encountered Note files created/modified Log ALL Errors: Every error goes in the plan file. This prevents repetition. ### Phase 3: Handle Errors Systematically The 3-Strike Protocol: ATTEMPT 1: Diagnose & Fix → Read error carefully → Identify root cause → Apply targeted fix ATTEMPT 2: Alternative Approach → Same error? Try different method → Different tool? Different library? → NEVER repeat exact same failing action ATTEMPT 3: Broader Rethink → Question assumptions → Search for solutions → Consider updating the plan AFTER 3 FAILURES: Escalate to User → Explain what you tried → Share the specific error → Ask for guidance Critical: if action_failed: next_action != same_action ### Phase 4: Verify Completion Use the 5-Question Reboot Test. If you can answer these, your context is solid: QuestionAnswer SourceWhere am I?Current phase in task_plan.mdWhere am I going?Remaining phasesWhat's the goal?Goal statement in planWhat have I learned?findings.mdWhat have I done?progress.md ### Quick Reference: Read vs Write SituationActionReasonJust wrote a fileDON'T readContent still in contextViewed image/PDFWrite findings NOWMultimodal → text before lostBrowser returned dataWrite to fileScreenshots don't persistStarting new phaseRead plan/findingsRe-orient if context staleError occurredRead relevant fileNeed current state to fixResuming after gapRead all planning filesRecover state ### Session Recovery When starting a new session, check for previous work: # Check if planning files exist ls task_plan.md findings.md progress.md 2>/dev/null # If they exist, read them all before continuing cat task_plan.md findings.md progress.md If planning files exist from a previous session: Read all three files to recover context Run git diff --stat to see what changed Update planning files with any missing context Continue from where you left off ### Templates Copy these to start: templates/task_plan.md — Phase tracking templates/findings.md — Research storage templates/progress.md — Session logging ### Scripts Helper scripts for automation: scripts/init-session.sh — Initialize all planning files scripts/check-complete.sh — Verify all phases complete ### References references/manus-principles.md — Context engineering principles from Manus ### Anti-Patterns Don'tDo InsteadUse TodoWrite for persistenceCreate task_plan.md fileState goals once and forgetRe-read plan before decisionsHide errors and retry silentlyLog errors to plan fileStuff everything in contextStore large content in filesStart executing immediatelyCreate plan file FIRSTRepeat failed actionsTrack attempts, mutate approachCreate files in skill directoryCreate files in your project ### NEVER Do NEVER start a complex task without task_plan.md — this is non-negotiable NEVER repeat a failed action exactly — track what you tried, mutate the approach NEVER ignore errors — log every error with resolution attempts NEVER rely on memory after >10 tool calls — re-read your plan NEVER skip the 2-Action Rule for visual content — multimodal data gets lost NEVER proceed past 3 failures without escalating — ask the user for help NEVER create planning files in the skill directory — they go in your project root ## Trust - Source: tencent - Verification: Indexed source record - Publisher: wpank - Version: 1.0.0 ## Source health - Status: healthy - Item download looks usable. - Yavira can redirect you to the upstream package for this item. - Health scope: item - Reason: direct_download_ok - Checked at: 2026-05-07T04:42:22.412Z - Expires at: 2026-05-14T04:42:22.412Z - Recommended action: Download for OpenClaw ## Links - [Detail page](https://openagent3.xyz/skills/planning-files) - [Send to Agent page](https://openagent3.xyz/skills/planning-files/agent) - [JSON manifest](https://openagent3.xyz/skills/planning-files/agent.json) - [Markdown brief](https://openagent3.xyz/skills/planning-files/agent.md) - [Download page](https://openagent3.xyz/downloads/planning-files)