SuperAPI AI Directory
Partner with us中文AI newsFavorites
Skillsdebug-task

debug-task

Testing & quality

View repository

Updated 10-15 13:06

About this skill

moon task debugger

A workflow-oriented diagnostic skill for troubleshooting moon tasks. This is not a reference manual — it guides you through a structured debugging flow so you can isolate the problem quickly.

For conceptual background, see the moon documentation.

Before you start: Ask the user for the <project>:<task> target to debug. If they haven't provided a specific target, prompt them for it — the diagnostic flow requires a concrete target to inspect.

Quick-start: 5-step diagnostic flow

Work through these steps in order. Most issues resolve by step 3.

Step 1: Inspect the resolved task configuration

The first thing to check is whether the task is configured the way the user expects. moon merges configuration from multiple sources (global tasks, project config, inheritance), so the resolved result can surprise people.

# Show the fully resolved task config (with inheritance applied)
moon task <project>:<task>

# Machine-readable version for programmatic inspection
moon task <project>:<task> --json

What to verify:

  • command vs script — if the command contains pipes (|), redirects (>), chained commands (&&), or complex syntax, it must use script, not command.
  • inputs — are they too broad (**/* captures everything) or too narrow (missing source files)? Check state.defaultInputs (true = using default **/*) and state.emptyInputs (true = explicitly set to [], or a task of the root-level project in a monorepo that configured no inputs). Both keys are omitted from the JSON entirely when false, as is state.setRunInCi.
  • outputs — are they declared for build tasks? Missing outputs means the cache can never hydrate artifacts. In v2.3+, outputs also affect the default cacheStrategy of any task that depends on this one (see Step 4).
  • toolchains — is the correct toolchain(s) assigned? An incorrect toolchain means wrong tool versions. v2.6+ The order is significant: configured toolchains come first, followed by related ones (those they require, and enabled toolchains that require them, like npm for node), and the first is the primary (used by $taskToolchain, and wins when toolchain environments are activated and paths are prepended to PATH).
  • deps — are task dependencies correct and complete? In v2.3+, each dep entry can carry a cacheStrategy (hash / ignored / outputs) that controls whether the dep contributes to this task's cache hash. If omitted, the default depends on whether the dep declares outputs. In v2.6+, each dep also has a type: required (default, runs before and must pass), cleanup (runs after the task, even when it fails), or wait (the task only waits for it to start). The human-readable moon task output labels the non-required ones, like db:stop (cleanup).
  • options — check persistent, runInCI, cache, affectedFiles, mutex, timeout, retryCount, allowFailure, expectFailure, envOverride, and os.
  • env — in v2.5+, environment variables can also be inherited from a workspace-level env in .moon/tasks/**/* (merged into the project's env, project wins), and the project can change the merge behavior via workspace.mergeStrategies.env. A variable with a surprising value may come from a layer outside the task. In v2.6+, toolchains that have been setup can also activate their environment (like JAVA_HOME), which sets variables the task config never mentions — but never overrides variables the task already configures. Also note that a task's env value only applies when the variable isn't already set in the shell — an exported shell value wins at runtime.
  • checks v2.4+ — shell scripts that run before the task. Their type determines the outcome: a requirement failing makes the task fail, all condition checks passing makes the task skip, and a fingerprint folds script output into the task hash. A surprising fail, skip, or cache invalidation often traces back to a check.
  • tags v2.3+ — labels for grouping tasks. Affects targets like :#quality and MQL taskTag queries. If a task isn't matched by a #tag target you expected, check this list.
  • type — an explicit type wins; otherwise build (has outputs), run (has a preset, or is persistent), else test. A persistent task with outputs is build.
  • preset — server or utility apply multiple option defaults at once.

Red flags:

  • command: 'eslint . && prettier --check .' — shell syntax in command is a task build error in v2 (task_builder::unsupported_command_syntax). Use script instead.
  • Empty outputs on a build task — cache will never restore artifacts.
  • inputs: ['**/*'] — too broad, cache invalidates on every change.
  • A non-persistent task with a required dep on a persistent task — a hard error when tasks are built. In v2.6+, use type: 'wait' to only wait for it to start — and since wait deps are still subject to the runInCI check (persistent and server preset tasks default to runInCI: false), also set runInCI: true on the persistent dep (or 'skip' on it, or false on the dependent), otherwise it fails with run_in_ci_mismatch.
  • A wait dep v2.6+ with no cleanup dep that stops it — a waited-on task is never stopped for you, so a non-persistent one keeps the pipeline running until it exits, and a persistent one keeps it running until moon is exited.
  • A wait dep on a task that has caching enabled v2.6+ — waited-on deps are started before the cache is checked, so on a cache hit the server starts and is then immediately stopped by its cleanup. Disable cache on tasks that wait on long-running deps.
  • command: 'noop' or nop / no-op — the task is intentionally a no-op and does nothing. moon treats these specially.
  • runInCI: 'only' — task runs in CI but NOT locally (common surprise).
  • runInCI: 'skip' — task is skipped in CI but relationships remain valid.
  • os set to a platform the user isn't on — the task is rewritten to a passing no-op at build time (moon task --json shows command: noop with cleared args/outputs).
  • allowFailure: true — the failure is still recorded and displayed, but the pipeline continues and moon exits successfully, so it's easy to miss.
  • expectFailure: true v2.6+ — the task passes when its command fails, and fails with "was expected to fail, but it passed" once the underlying problem is fixed (remove the setting). It's never cached, and exit codes 126/127 or a signal don't count as the expected failure.
  • A condition check present v2.4+ — the task will skip whenever all conditions pass. A task that "never runs" may have a condition that always passes.
  • A fingerprint check present v2.4+ — its script output is hashed, so volatile output (timestamps, versions) causes cache misses on every run.

Step 2: Run with maximum verbosity

If the config looks right, run the task with debug logging to see what moon is actually doing under the hood.

# Debug-level logging with cache bypass (note: --force also skips affected checks)
moon run <project>:<task> --log debug --force

# Deep debugging: reveal env vars and stdin passed to the process
MOON_DEBUG_PROCESS_ENV=true MOON_DEBUG_PROCESS_INPUT=true moon run <project>:<task> --log trace --force

What to look for in the logs:

  • Toolchain resolution — is the right version of node/deno/bun/etc being used?
  • Hash generation — what sources are being hashed?
  • Affected status — is the task being skipped because it's "not affected"? --force runs tasks even when they're not affected, so drop it (and pass --affected) when diagnosing this.
  • Process execution — what command is actually being spawned?

Visualize the execution graph to spot dependency issues:

moon action-graph <project>:<task> --dot   # DOT format
moon action-graph <project>:<task> --json  # JSON format

Always pass --dot or --json as an agent. Without them, the command starts a local web server and opens a browser, and blocks until it's interrupted.

For all graph commands and output formats, see references/environment-debug.md.

Step 3: Inspect cache state

If the task runs but produces wrong results, or runs when it shouldn't, or doesn't run when it should, the cache is the likely culprit.

# Inspect a hash manifest to see what inputs were hashed
moon hash <hash>

# Compare two hashes to see what changed between runs
moon hash <hash1> <hash2>

# Short-form hashes work too
moon hash 0b55b234 2388552f

v2.6+ Hash manifests are stored as blobs in the local content-addressable cache (.moon/cache/blobs/), not as .moon/cache/hashes/<hash>.json files, so always read them through moon hash. If it reports "Unable to find a hash manifest", the manifest may have been garbage collected — re-run the task with --force to regenerate it.

For cache file locations, hash interpretation, and the --force vs --cache off comparison, see references/cache-issues.md.

Step 4: Diagnose the problem type

Use this table to jump to the right reference:

Symptom Likely cause Quick check Reference
Task doesn't exist Inheritance not applied — check inheritedBy conditions in .moon/tasks/**/* against the project's toolchains, config.language, config.stack, config.layer, config.tags via moon project <name> --json (a detected-only language doesn't match inheritedBy.languages) moon task <target> --json references/config-mistakes.md
"Nothing to do" --affected + no changes, runInCI: false, or inheritedBy mismatch (global task not inherited) Check flags, options.runInCI, and inheritedBy references/decision-tree.md
--affected misses changed files (or runs everything) Shallow git clone in CI — with no base, moon warns "Detected a shallow checkout…" and disables affected (so everything runs); with an explicit base, diffs may be inaccurate Check clone depth; use full history (fetch-depth: 0) references/decision-tree.md
Task fails: "…is unable to run as the requirement check … failed" v2.4+ A requirement check script exited non-zero (task_runner::requirement_check_failed); a failing fingerprint script also fails the task (task_runner::hash_check_failed) moon task <target> --json — inspect checks references/config-mistakes.md
Task skipped, not affected/CI-related v2.4+ All condition checks passed, so the task was intentionally skipped moon run <target> --log debug — look for "Skipping task as all conditional checks have passed" references/config-mistakes.md
Task errors on execution Wrong command/script, bad toolchain moon run <target> --log debug references/config-mistakes.md
Stale cache (cached when it shouldn't be) Inputs too narrow, missing env vars, or dep cacheStrategy: 'ignored' (the v2.3 default for output-less deps) moon hash <hash> references/cache-issues.md
Cache miss (re-runs every time) Inputs too broad, volatile outputs, or dep cacheStrategy: 'hash' propagating upstream churn moon hash <h1> <h2> references/cache-issues.md
Cache miss from a fingerprint check v2.4+ A fingerprint check's script output is volatile (timestamps, PIDs), changing the hash every run moon hash <h1> <h2> — look for the check hash references/cache-issues.md
Outputs not restored after cache hit outputs misconfigured; or v2.5+ a daemon-side storage failure (only with the daemon and casOutputsCache or a remote cache) — logged only by the daemon, and a failed hydrate becomes a silent cache miss Check .moon/cache/outputs/; moon daemon logs references/cache-issues.md
Task fails with missing_outputs after remote cache errors A transient HTTP remote cache error (like a 503) dropped some blobs while hydrating; v2.6+ requests are retried (remote.cache.retryCount, default 3) moon run <target> --log debug — look for "Failed to retrieve blobs" references/cache-issues.md
Env var has unexpected value v2.5+ Workspace-level env in .moon/tasks/**/* merged in, or workspace.mergeStrategies.env changed the merge behavior; or the variable is already set in the shell (a task's env, and v2.6+ deps.*.env, only applies when it isn't, unless v2.6+ options.envOverride is enabled) moon task <target> --json — inspect env; env | grep <VAR> references/config-mistakes.md
Cache behaves differently across git worktrees v2.5+ cache.sharedWorktreeCache (alias unstable_sharedWorktreeCache) shares blobs/manifests via the base checkout's .moon/cache Check the setting (both names) and MOON_CACHE_SHARED_WORKTREE_CACHE references/cache-issues.md
New dependency cycle error after upgrading to v2.5 Async graph building (now default) validates cycles strictly, per dependency-scope partition Set experiments.asyncGraphBuilding: false to confirm references/decision-tree.md
Build re-runs on every upstream input change v2.3+ Dep using default cacheStrategy: 'hash' instead of 'outputs' moon task <target> --json — inspect dep entries references/cache-issues.md
Task not matched by #tag target v2.3+ Missing tags on the task, or mergeTags dropped them during inheritance moon task <target> --json — check tags references/config-mistakes.md
Task hangs / pipeline stuck v2.6+ A wait dep that is never stopped by a cleanup dep, a persistent task that was requested (it runs until interrupted, by design), or a persistent/wait task holding a mutex another task needs moon action-graph <target> --dot; moon task <target> for dep types references/config-mistakes.md
Cleanup dep didn't run v2.6+ The task was hydrated from the cache or skipped (nothing to clean up), the pipeline was interrupted by a signal (Ctrl+C), or the cleanup has runInCI disabled moon run <target> --log debug — look for "nothing to clean up" references/config-mistakes.md
Task can't reach a wait dep (connection refused) v2.6+ wait only waits for the dep to start, not to be ready Poll a health check before sending requests references/config-mistakes.md
Persistent task doesn't run in CI v2.6+ Persistent tasks default to runInCI: false; when enabled, they only run for the tasks that depend on them, or when explicitly targeted moon task <target> --json — inspect options.runInCI references/config-mistakes.md
Output not displayed (or displayed when it shouldn't be) outputStyle applies to transitive targets only; v2.4.3+ primary targets always display output (except v2.5.2–v2.5.x in CI or on cache hits), unless v2.6+ explicitTaskOutputStyle is enabled or --output-style is passed Check options.outputStyle and the command line references/config-mistakes.md
Tasks re-run once after upgrading to v2.6 A task's toolchains list (part of its hash) is now ordered by configuration instead of hash order, which invalidates the cache once for tasks with multiple resolved toolchains Expected; the second run should hit the cache references/cache-issues.md
moon hash can't find a manifest v2.6+ Hash manifests are blobs in the local CAS, and may have been garbage collected moon run <target> --force, then moon hash the new hash references/cache-issues.md
Env var set that the config doesn't mention v2.6+ A toolchain activated its environment (like JAVA_HOME) MOON_DEBUG_PROCESS_ENV=true moon run <target> --log debug references/environment-debug.md
Task is slow Dep chain bottleneck, no parallelism moon action-graph <target> --dot references/decision-tree.md
Task does nothing (no-op) Command is noop/nop/no-op moon task <target> --json references/config-mistakes.md
Task "fails silently" allowFailure: true — the task is still marked failed, but the pipeline continues and the exit code ignores it Check options.allowFailure references/config-mistakes.md
Task fails: "…was expected to fail, but it passed" v2.6+ options.expectFailure is set and the command now succeeds, so the known failure was fixed (task_runner::unexpected_pass). Exit codes 126/127, signals, and timeouts are not the expected failure (task_runner::unexpected_failure) Remove options.expectFailure (or fix the command if it exited 126/127) references/config-mistakes.md
Task skipped locally runInCI: 'only' set Check options.runInCI references/config-mistakes.md
Task skipped in CI runInCI: false or 'skip' Check options.runInCI references/config-mistakes.md
Mutex contention / hang Tasks share the same mutex; a persistent or wait task holding it never releases it Check options.mutex references/config-mistakes.md
Task times out timeout option set too low Check options.timeout references/config-mistakes.md

Step 5: Validate the fix

After making changes, verify the fix actually worked:

# Bypass cache to force a fresh run
moon run <project>:<task> --force

# Disable cache entirely (no reads OR writes)
moon run <project>:<task> --cache off

# Verify the resolved config reflects your changes
moon task <project>:<task> --json

--force vs --cache off:

  • --force ignores existing cache but writes new cache after execution — and also runs tasks that aren't affected (it skips affected checks).
  • --cache off disables caching entirely — no reads, no writes.

For all cache modes, see references/cache-issues.md.

Common mistakes at a glance

These are the issues that come up most often. For details and fixes, see references/config-mistakes.md.

  • Shell syntax in command — pipes, &&, redirects require script; v2 rejects these as parse errors.
  • Missing outputs on build tasks — cache can never hydrate artifacts.
  • Overly broad inputs — **/* invalidates cache on every change; be specific.
  • Volatile outputs — timestamps or absolute paths in build artifacts cause permanent cache misses.
  • Persistent task in deps — a required dep on a persistent task is a hard error; tasks named dev/start/serve auto-get the server preset. In v2.6+, depend on it with type: 'wait', enable runInCI on it (it defaults to off, and wait deps are still CI-checked), and pair it with a cleanup dep that stops it.
  • --affected vs --force confusion — --affected restricts which tasks run; --force bypasses the cache and skips affected checks, so --force overrides --affected (tasks run even when not affected).
  • allowFailure: true hiding errors — the failure is still recorded and displayed, but the pipeline continues and moon exits successfully; check stderr at .moon/cache/states/<project>/<task>/stderr.log.
  • mutex contention — shared mutex serializes tasks; a persistent (or wait) task holding a mutex never releases it, so other tasks with that mutex hang.
  • runInCI: 'only' — task silently skips when run locally (most surprising variant).
  • Missing outputs flip dep cacheStrategy v2.3+ — a dep without outputs now defaults to cacheStrategy: 'ignored'. Downstream tasks stop invalidating on its changes; set cacheStrategy: 'hash' explicitly to restore the pre-v2.3 default.
  • MQL tag fields on task queries — the query is a positional argument (moon query tasks "taskTag=quality"; there's no --query flag on moon query). In v2.3–v2.4, taskTag= silently matched nothing on task queries (this also broke task tag glob targets like :#tag-*); v2.5 fixed taskTag. projectTag=/tag= (and other project fields, except project=) still match nothing on task queries — query projects instead. On older versions, filter task tags with the --tags flag (moon query tasks --tags quality).
  • A checks script silently changes task behavior v2.4+ — a requirement failing aborts the task, a passing condition skips it, and a fingerprint mixes script output into the hash. Inspect checks in moon task <target> --json when a task fails, skips, or re-runs for no obvious reason.
  • Shallow git clone breaks --affected — with a shallow clone (depth 1) and no explicit base, moon warns "Detected a shallow checkout…" and disables affected filtering for moon run/ci/ exec, so everything runs (moon query … --affected gets an empty list instead). With an explicit base (--base, MOON_BASE, or a detected PR base), it diffs anyway and the result may be inaccurate. Use a full clone (fetch-depth: 0), or a blobless partial clone with full history (git clone --filter=blob:none).
  • Experiments are now on by default v2.5+ — asyncGraphBuilding, asyncAffectedTracking, and nativeFileHashing default to enabled. When bisecting graph, affected, or hashing oddities, disable the relevant experiment (config or MOON_EXPERIMENT_*=false) and compare — but also check the user's shell/CI for MOON_EXPERIMENT_* or MOON_CACHE_* overrides that silently change behavior.
  • Daemon archiving/hydration failures are invisible in the main process v2.5+ — when the daemon is enabled and outputs go to the local CAS (casOutputsCache) or a remote cache, they're archived and hydrated through the daemon, and storage failures only appear in moon daemon logs (default .tar.gz archives are always handled in-process). To rule the daemon out, re-run with MOON_DAEMON=false.
  • Workspace-level env is a new inheritance layer v2.5+ — .moon/tasks/**/* files can define env inherited by all matching projects. Project values win on conflict, unless workspace.mergeStrategies.env says otherwise (append, prepend, preserve, replace).
  • Expecting a cleanup dep to always run v2.6+ — cleanups are skipped when there's nothing to clean up (the task was hydrated from the cache or skipped, and didn't start a wait dep), and they never run when the pipeline is interrupted by a signal like Ctrl+C.
  • cleanup and wait deps don't affect the cache v2.6+ — they never contribute to the task's hash (their cacheStrategy can only be ignored), and they always run their command instead of being hydrated.
  • Persistent tasks run within the pipeline v2.6+ — they no longer wait until every other task has finished; they start as soon as their own deps complete. In CI, they're skipped unless runInCI is explicitly enabled, and even then only run for their dependents or when explicitly targeted. moon check skips them unless another task depends on them.

When to load references

Each reference file covers a specific problem domain in depth. Load them only when the diagnostic flow points you there — don't load everything upfront.

Reference When to load
references/decision-tree.md When the symptom doesn't match the quick table above, or you need a systematic walk-through of all possibilities.
references/cache-issues.md When the problem is clearly cache-related: unexpected hits, unexpected misses, outputs not restoring.
references/config-mistakes.md When the task config is wrong: command vs script, inheritance bugs, presets, persistent tasks, dependency types (cleanup/wait), affectedFiles, mutex, timeout, retries, runInCI variants, allowFailure, os, outputStyle.
references/environment-debug.md When you need to go deeper with env vars, log levels, trace profiles, or inspection tools.