forjd

Synthesize startup customer interview notes, outreach replies, validation calls, and test results into evidence, contradictions, signal strength, and a proceed/park/kill recommendation before v1 scoping, with optional formatted single-file HTML artifact output.

Stress-test a captured startup problem with sharp questions before validation or product scoping, with optional formatted single-file HTML artifact output.

Generate a raw inventory of startup problem candidates from founder unfair access, domain observations, workflows, complaints, and recent enabling changes before logging or validating any one idea, with optional formatted single-file HTML artifact output.

Capture raw startup pains and sharpen them into testable problem statements before proposing products, with optional formatted single-file HTML artifact output.

Scope a startup idea into a narrow wedge and 2–4 week v1 that tests the riskiest assumption, with optional formatted single-file HTML artifact output.

Design small credible demand tests, interviews, smoke tests, and kill criteria before building a startup v1, with optional formatted single-file HTML artifact output.

Browser automation CLI for AI agents. Use when the user needs to interact with a web page or web app in a real browser, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting page data, testing rendered UI, or running browser-based health checks. Do not use for code review, API-only checks, generic QA, or general web research unless browser interaction is required.

Create standardised GitHub pull requests using the gh CLI. Use when the user wants to create a PR, open a pull request, submit changes for review, or says things like "PR this", "open a PR", or "submit for review". Enforces conventional commit titles, structured body templates, labels, and reviewers.

Action GitHub pull request review feedback. Use when the user wants to address PR review comments, respond to code review feedback, fix review issues, or says things like "action the review", "address feedback", "handle review comments", or "go through the PR comments". Fetches all unresolved review threads, triages each as actionable or not, implements fixes, and replies with rationale where not actioning.

Audit and harden GitHub repository security settings using the gh CLI. Use when the user wants to review or improve repository security posture, enforce branch protection, enable secret scanning, configure merge policies, lock down GitHub Actions permissions, or apply security best practices. Triggers on requests to "harden", "secure", "lock down", or "audit" a GitHub repository, even if they just say "make this repo more secure".

Audit implemented Signal Box plans against repository code and docs. Use after a plan or phase has been implemented to compare implementation against docs/plans acceptance criteria, identify missed scope, check out-of-scope leakage, and update plan frontmatter status to verified, needs-work, or implemented.

Implement Signal Box planning documents. Use when Codex is asked to build a phase or execute a plan from docs/plans, including updating plan frontmatter from planned to in-progress to implemented, keeping work scoped to the selected plan, updating documentation, and preparing the result for audit.

Create or update Signal Box planning documents. Use when writing docs/plans phase files, alpha roadmaps, implementation plans, scope-to-plan breakdowns, or checking that plans follow the repository's standard planning structure and stay aligned with docs/scope.md.

Audit and manage Tilezo's docs/IDEAS.md statuses. Use when the user asks whether ideas have been implemented, wants implemented ideas marked off, wants partial ideas separated from complete work, or wants caveats and next steps captured in FOLLOW_UPS.md.

Capture and append product, design, engineering, gameplay, UX, or planning ideas to Tilezo's docs/IDEAS.md file. Use when the user asks to note down, jot down, capture, record, save, add, or remember an idea, brainstorm item, future feature, TODO-like concept, or rough thought for the project.

Design, generate, validate, and integrate layered isometric avatar sprites for Tilezo. Use when working on Habbo-like character assets, avatar spritesheets, avatar layer manifests, character appearance options, PixiJS avatar rendering, or replacing procedural Graphics avatars with proper layered sprites.

Browser automation CLI for AI agents. Use when the user needs to interact with websites, including navigating pages, filling forms, clicking buttons, taking screenshots, extracting data, testing web apps, or automating any browser task. Triggers include: any mention of 'browse', 'check the page', 'take a screenshot', 'test the UI', 'fill the form', 'click the button', 'QA', 'visual check', 'healthcheck', and any task requiring a real browser.

Create plugins for the browse CLI. Use when the user wants to extend browse with custom commands, add lifecycle hooks, or build a browse plugin. Triggers include: 'create a plugin', 'add a command to browse', 'extend browse', 'browse plugin', 'hook into browse'.

Rewrite, draft, and review prose so it is clear, specific, human, and appropriate to context. Use this skill when improving emails, essays, documents, reports, UI copy, marketing copy, posts, or any text that may sound generic, AI-written, over-polished, verbose, salesy, evasive, or structurally formulaic. Supports voice calibration, anti-slop audits, context-aware taste decisions, and final pre-flight checks.

Capture and recall working context with the `aid` CLI in this repository. Use when resuming prior work, recording important findings, tracking meaningful tasks, storing engineering decisions, searching saved context, or generating a handoff for the next session.

Use when needing to find relevant git commits by meaning — investigating what changed, finding related work, understanding history, or locating commits about a feature/bug/refactor. Requires @forjd/git-search to be installed.

Decision guide for choosing between inline work, one subagent, parallel subagents, and workflow orchestration. Use when planning a task that could involve searching many files, independent parallel strands, large-scale fan-out, specialized agents, or when tempted to spawn agents, so the agent picks the cheapest level that fits.

Render seven-segment LCD text in terminal, SVG, or PNG using the seg-lcd-rust CLI.

Use this skill when working in a software repository and you need to create, update, preview, or verify deterministic agent context with Minimap in AGENTS.md, CLAUDE.md, or another agent instruction file.

Use when updating, reviewing, prioritizing, or turning ideas into tasks for this repository's ROADMAP.md. Applies to roadmap status markers, dependency-aware ordering, scope clarification, completion criteria, and keeping README.md or CHANGELOG.md aligned with shipped roadmap work.