qtsone/learning / skills/tutor
tutor
FlaggedTurn this agent into a personal programming tutor with a versioned 0-to-expert curriculum (Go available; more languages coming). Use when the user wants to learn a programming language, start or resume lessons, get their exercise code graded, check learning progress, or contribute curriculum fixes — e.g. "/tutor go", "/tutor resume", "teach me Go", "continue my lessons", "grade my exercise".
curriculum/content/focus/cli-tooling/terminal-ux/quiz.json
View on GitHub{
"pass_rule": "all core questions substantially correct in the learner's own words; stretch questions inform the grade, not the gate",
"questions": [
{
"id": "two-streams",
"difficulty": "core",
"prompt": "I run your tool as `tool | jq .` and as `tool > out.txt`. What goes to stdout, what goes to stderr, and how do you decide which is which?",
"expected_points": [
"stdout carries the data the user asked for — what flows down a pipeline",
"stderr carries diagnostics: progress, prompts, warnings, verbose logs",
"the test is 'would a script parse this?', not 'is it bad news?'",
"progress or prompts on stdout corrupt piped data; results on stderr starve the pipeline",
"the exit code is the third channel"
]
},
{
"id": "tty-detection",
"difficulty": "core",
"prompt": "How does your program know whether it is talking to a terminal, and what changes when it isn't? Where in the program does that question get asked?",
"expected_points": [
"stat the file and test os.ModeCharDevice, or golang.org/x/term.IsTerminal for the accurate answer",
"caveat: /dev/null is a character device too; Windows consoles need extra setup",
"when it isn't a terminal: no colour, no cursor redraws, no prompting — the information survives, the decoration doesn't",
"detection happens once at the edge and is passed down as a bool, which is what makes everything below testable",
"stdin, stdout and stderr can each be redirected independently"
]
},
{
"id": "color-precedence",
"difficulty": "core",
"prompt": "State your colour decision chain from the top. What does NO_COLOR do, and what happens when NO_COLOR=1 and --color=always are both present?",
"expected_points": [
"an ordered, written-down chain: never / always / NO_COLOR / TERM=dumb / is-a-terminal",
"NO_COLOR is a cross-tool convention: set and non-empty means no colour, whatever the value",
"NO_COLOR set to the empty string is not a veto",
"the explicit per-invocation flag beats the environment default here — and the learner can defend that ordering",
"the chain is a pure function of (mode, isTerminal, env), so it is testable without a terminal",
"--no-color is a shorthand folding into the same chain as --color=never, not a second mechanism: users expect the boolean, the three-valued --color is what makes 'auto' expressible, and both must land on one decision function"
]
},
{
"id": "progress-degradation",
"difficulty": "core",
"prompt": "Describe what your progress indicator writes on a terminal and what it writes when stderr is a CI log file. Why that stream, and why is it driven by Update calls rather than a timer?",
"expected_points": [
"on a terminal: \\r, the status, and \\x1b[K to erase the tail of a longer previous line, redrawn in place",
"off a terminal: nothing during the run, one plain summary line at the end — no \\r, no escapes",
"it goes to stderr so redirected or piped stdout stays clean data",
"explicit updates keep the output deterministic and testable; a timer means a goroutine, wall-clock flakiness, and races",
"real tools throttle redraws, but that belongs outside the tested piece"
]
},
{
"id": "prompt-contract",
"difficulty": "core",
"prompt": "Walk me through your Ask: the default, validation, and what happens in a CI job with no terminal and no default.",
"expected_points": [
"the default is shown in brackets and an empty answer takes it",
"validation rejects with a message and re-asks, with a cap so the loop terminates",
"with no terminal it never reads or writes: the default, or ErrNotInteractive naming what to pass instead",
"a tool that blocks on a prompt in a pipeline or CI is broken; --yes exists for this",
"one bufio.Scanner for the session — a fresh one per question eats buffered input"
]
},
{
"id": "tui-tradeoff",
"difficulty": "core",
"prompt": "What does a TUI framework like Bubble Tea give you that line-based output doesn't, and when is it the wrong answer?",
"expected_points": [
"the Elm-style loop: a model, Update(msg) returning a new model plus commands, View rendering the whole screen",
"the framework owns the event loop, alternate screen, raw input and repainting",
"right for long-lived interactive sessions the user stays inside",
"wrong for run-do-work-report tools: not pipeable, greppable or scriptable, awkward in CI, over ssh, in dumb terminals and for screen readers",
"and markedly harder to test than a function writing to an io.Writer"
]
},
{
"id": "json-contract",
"difficulty": "stretch",
"prompt": "You want to realign the human output columns and rename a JSON field. Which one can you do in a patch release, and why?",
"expected_points": [
"the human format is not an API; the machine format is",
"JSON shape changes like an API changes — other people's scripts depend on it",
"JSON on stdout alone, diagnostics on stderr, never styled",
"emit a document like {\"results\": …} rather than a bare array so it can grow fields",
"encode an empty slice, not a nil one, so consumers never meet null"
]
},
{
"id": "alignment",
"difficulty": "stretch",
"prompt": "Why does `fmt.Sprintf(\"%-4s\", painted)` break your columns when colour is on, and what did you do instead?",
"expected_points": [
"%-4s pads to the length of the string it is given, and the escape characters count toward that length even though they occupy no columns",
"coloured lines end up shorter on screen than plain ones",
"pad by the visible status word, then paint — or paint and pad separately"
]
}
]
}
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |