Humungousaur

Design, implement, organize, document, and verify Humungousaur platform OS collectors for macOS, Windows, or Linux using official OS API references, shared event envelopes, helper health, privacy-first metadata, and repo-native adapter boundaries.

Parent skill for code inspection, debugging, tests, CI, reviews, dependency security, skill authoring, experiments, and engineering workflows.

Apply user, stranger, group, channel, and action access tiers using native channel policy, allowlists, approvals, and cognitive review.

Assess the assistant's readiness, capability fit, safety risks, tool coverage, and verification plan before or during complex work. Use when work is high-impact, ambiguous, risky, long-running, or failure-prone.

Review recent agent behavior, uncertainty, mistakes, interaction quality, and lessons so the assistant can improve future work. Use after complex tasks, failures, user corrections, or long sessions.

Treat unmentioned group-room chatter as quiet context unless policy, mention, direct request, or model-led cognition decides a response is appropriate. Use for Slack channels, Discord servers, Telegram groups, Matrix rooms, and similar shared spaces.

Require explicit human approval before purchases, messages, installs, file writes, remote API mutations, app control, or other irreversible actions.

Inspect run history, approvals, outbox, interpreter manifests, screenshots, and memory events to reconstruct what the agent actually did.

Run bounded autonomous cycles over queued events, due wakeups, ready tasks, and optional initiative while preserving approvals and stop conditions. Use when configuring, testing, or operating Humungousaur as a continuing assistant.

Prevent bot-to-bot reply loops by detecting bot-authored inbound messages, pair suppression, channel policy, and ambient context rules. Use when handling channels that may receive messages from other assistants or automation.

Audit Humungousaur's effective tools, skills, plugins, channels, providers, local services, and missing capability gaps. Use before large integrations, after adding tools, or when the user asks what the agent can currently do.

Curate assistant memory by retaining useful knowledge, summarizing noisy history, archiving stale records, and forgetting unsupported details. Use when memory feels noisy, outdated, duplicated, or too large.

Update assistant persona, communication style, user preferences, boundaries, and stable facts from evidence. Use when the user corrects behavior, states preferences, or asks the assistant to become more personally useful.

Screen untrusted text, webpages, documents, and channel messages for prompt-injection, exfiltration, tool-abuse, and instruction-conflict risks.

End a work session by summarizing completed work, blockers, tests, pending commitments, memory lessons, and optional git status. Use when the user asks to wrap up, checkpoint, commit, or preserve learnings.

Parent skill for core agent cognition, safety, memory, taskflow, approvals, system readiness, and self-improvement workflows.

Check local system readiness, disks, environment, integrations, voice/browser/channel providers, and dependency health before or during agent workflows.

Track tasks, commitments, blockers, statuses, and next actions across ongoing user work. Use when the user asks what is pending, what changed, what is blocked, what should be updated, or how to keep a task moving.

Coordinate durable multi-step work with goals, tasks, waits, wakeups, autonomous cycles, and child/delegated tasks.

Create, review, evaluate, and cancel future reminders, wakeups, and triggers for the assistant. Use when the user wants follow-ups, timed checks, reminders, monitors, or future task continuation.

Use browser, Playwright, and computer-use style tools through observed state and approval-gated actions.

Use search, static fetch, and live browser tools for current web evidence with observation-first browser discipline, date verification, anti-churn rules, and approval-gated interactions.

Use Browser Use-derived capabilities through Humungousaur's native browser tools first, then approval-gated Browser Use autonomous delegation when native interaction is insufficient.

Design and implement polished, domain-appropriate frontend interfaces using existing app conventions, visual verification, and native browser testing.

Test live browser apps with Humungousaur's Playwright-backed browser tools, observation-first actions, screenshots, downloads, PDFs, and evidence-backed completion checks.

Parent skill for web browsing, browser automation, web evidence gathering, web forms, browser testing, extraction, and web artifact workflows.

Extract structured information from web pages with source provenance, browser observations, and summaries while respecting copyright, privacy, and tool boundaries.

Fill, review, and submit browser forms safely using observed fields, explicit values, approval gates, and verification after each browser action.

Operate Humungousaur WebChat-style browser conversations through native channel manifests, WebSocket/runtime contracts, outbox preparation, and interaction-harness routing.

Prepare shopping, checkout, and payment decisions without executing purchases or financial actions unless explicitly approved and verified.

Native Humungousaur skill for Shop App. Use when a task calls for shop app workflows, readiness checks, artifacts, or approval-gated local/provider actions.

Native Humungousaur skill for Shopify. Use when a task calls for shopify workflows, readiness checks, artifacts, or approval-gated local/provider actions.

Research and compare products, prices, constraints, and purchase options while keeping checkout, payments, and messages approval-gated.

Parent skill for shopping, payments, maps, rail, flight, and travel booking research with approval-gated purchase and booking boundaries.

Plan routes, places, trips, and local itineraries from current evidence while marking missing native map/API adapters and avoiding unapproved bookings.

Native Humungousaur skill for Agentmail. Use when a task calls for agentmail workflows, readiness checks, artifacts, or approval-gated local/provider actions.

Normalize chat-channel setup, inbound events, outbound preparation, ambient context, and bot-loop policy through Humungousaur's native Gateway tools. Use when working across multiple messaging channels or onboarding a new communication surface.

Operate Humungousaur Gateway chat channels with setup checks, policy gates, ambient context, bot-loop safety, and audited outbound delivery.

Read provided email evidence, draft and summarize email, prepare follow-ups, and identify native adapter gaps for Gmail, Outlook, IMAP, SMTP, or local mail workflows.

Native Humungousaur skill for Himalaya. Use when a task calls for himalaya workflows, readiness checks, artifacts, or approval-gated local/provider actions.