molotroka123-cell

video-editing

by molotroka123-cell

Edit owner-attached local media through the shared Video Studio project commands, preview, export and independent verification.

solana-volume-suite

by molotroka123-cell

Autonomous Solana AI Market Making, volume generation, Pump.fun bonding curve progression, Raydium AMM / Jupiter v6 swaps, Zero-Knowledge AES-256-GCM sub-wallet management, Jito MEV sandwich protection, and real-time Command Center telemetry at http://localhost:8501.

compact

by molotroka123-cell

external-evidence-check

by molotroka123-cell

Verify a claim taken from outside the system — a web page, a search result, an MCP server, a model's own recall — before it is written into long-term memory or used to justify a decision.

requirement-clarify

by molotroka123-cell

Resolve an ambiguous requirement before any work starts, by separating what was actually stated from what is being assumed, and either asking one blocking question or proceeding under a written assumption.

skill-evaluator

by molotroka123-cell

Evaluate a new or modified skill against its previous version on reproducible tasks before promotion, preventing self-improvement from silently degrading quality, safety or cost.

context-builder

by molotroka123-cell

Assemble the smallest high-quality task context from memory, semantic code search, current files, tool results and user instructions while protecting the model context window from irrelevant material.

permission-auditor

by molotroka123-cell

Review an agent, skill, MCP server or workflow permission set for unnecessary access, privilege expansion and dangerous combinations before it is enabled.

proof-before-done

by molotroka123-cell

Require concrete runtime evidence before an agent, reviewer or mission marks work complete, preventing UI-only, mock-only or self-reported success from becoming false DONE status.

memory-recall

by molotroka123-cell

Retrieve only the relevant prior decisions, lessons, conventions and project history needed for the current task, using progressive local memory search instead of loading whole histories.

failure-retrospective

by molotroka123-cell

Analyze a failed or repeatedly struggling task, identify the actual failure class, extract reusable lessons and propose bounded changes to routing, memory, skills or tooling.

memory-curator

by molotroka123-cell

Convert completed work into concise durable memory containing decisions, lessons, successful procedures, failures worth remembering and next actions without storing raw transcript noise.

obsidian-knowledge

by molotroka123-cell

Use the configured Obsidian vault as a read-mostly local knowledge base, with agent writes restricted to the BOSSMAN Memory folder.

night-mission

by molotroka123-cell

Convert a long autonomous goal into a persistent overnight mission with checkpoints, bounded retries, resource and cloud budgets, review gates, and a clear morning handoff.

browser-research

by molotroka123-cell

Navigate and inspect websites efficiently using DOM-first browser automation, screenshots only when visual understanding is needed, and strict approval gates for sensitive actions.

openrouter-sync

by molotroka123-cell

Synchronize OpenRouter models into BOSSMAN with pricing, context, modalities, capabilities and live compatibility probes, while preserving user aliases and routing policies.

repo-audit

by molotroka123-cell

Audit a software repository before changes: map architecture, identify real vs mock functionality, tests, risks, technical debt, and the smallest safe implementation path.

model-eval

by molotroka123-cell

Benchmark and validate a local or cloud model for real BOSSMAN roles, measuring speed, memory, capabilities, tool use, context behavior, stability, and cost rather than relying only on public benchmarks.

safe-terminal

by molotroka123-cell

Operate a terminal for coding and system tasks with scoped working directories, command-level permissions, visible logs, timeouts, and no silent privilege escalation.

safe-code-change

by molotroka123-cell

Implement a code change safely in an isolated branch/worktree, verify it with focused tests and browser QA when relevant, and prepare a clean commit without unsafe repository operations.

skill-forge

by molotroka123-cell

Create or improve reusable agent skills from repeated successful workflows and explicit user requirements, keeping them portable across BOSSMAN, OpenCode and Claude-compatible skill directories.