agentleFS
dougcunhadougcunha/TokenHound/.agents/skills/no-workarounds/SKILL.md

no-workarounds

Skill3 starsChanged 36 days agoMITView on GitHub

No workarounds. Use when debugging code or test failures, planning a fix, or reviewing changes that may hide a defect, including .NET/C# nullability, async, dependency injection, and EF Core issues. Not for formatting- or docs-only edits.

3 other projects use a version of this.

Adopted by

Most recently taken up by

The file

We split this skill into the rules it gives an agent

5 rules. 0 are common, 5 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 3 edited versions we read keep these.

  • 5

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use these signals to investigate a possible workaround.

    line 30
  2. Distinctive:

    Not every root cause is yours to fix.

    line 50
  3. Distinctive:

    When all four hold, contain it:

    line 59
  4. Distinctive:

    If any condition fails, fix the root cause.

    line 69
  5. Distinctive:

    When discussing the rationale or pressure to skip a root-cause fix, read references/philosophical-foundations.md.

    line 73

“Common” is measured: kept by at least 6 in 10 of the 3 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

No Workarounds

A workaround makes a failure disappear while leaving its cause intact. Fix the source, not the signal. Judge the violated contract and the evidence, not syntax alone: casts, null checks, retries, and framework initialization can be legitimate. Apply this distinction when reading the catalogs too.

The gate: run before any fix

1. Capture the failing behavior or diagnostic and the expected contract.
2. Trace the relevant data flow, lifecycle, or configuration to a cause supported by evidence.
3. Check that the proposed fix repairs that cause while preserving valid behavior.

Silencing a signal -> redesign the fix against the root cause.
Cause still unknown -> gather evidence; do not present a hypothesis as a verified fix.
Root cause is external -> evaluate the escape valve.

Complete the fix when the original reproduction passes, relevant regression checks preserve the contract, and no remaining suppression or fallback conceals the defect. For a plan or review, identify the causal evidence and the checks required to verify the proposed fix; distinguish these from checks actually run.

The seven signals

Use these signals to investigate a possible workaround. Confirm the defect before changing a valid language or framework idiom.

CategorySuspected hidden defectFix the source by
TYPE: forced assertions, dynamic, unjustified !Types conceal invalid or absent dataCorrecting the contract, initialization, or boundary validation
LINT: diagnostic suppression, disabled analyzersA real warning is hiddenFixing the finding; for a proven false positive, documenting a narrowly scoped suppression instead of disabling checks broadly
SWALLOW: empty catch, catch-and-defaultFailure becomes apparent successHandling expected errors explicitly and propagating unexpected failures; preserving context at the responsible error boundary
TIMING: arbitrary delay, blocking async, blind retriesOrdering or ownership is wrongAwaiting completion or coordinating readiness; retrying only identified transient failures with bounded, cancellation-aware policies
PATCH: global or library-internal mutationA missing API contract is bypassedUsing a supported extension point, adapter, or upstream correction
SCATTER: fallback chains for required dataInvalid input escapes its boundaryValidating required data at entry; retaining null handling for genuinely optional values
CLONE: copy-and-tweakA forced abstraction spreads the defectSharing behavior with the same contract, or separating distinct responsibilities

Select the relevant reference

  • For .NET/C# fixes or reviews, read references/dotnet-csharp.md in full before choosing the fix. It covers runtime and SDK compatibility, nullable contracts, async, DI, EF Core, configuration, and verification.
  • For JavaScript/TypeScript signals, read references/workaround-catalog.md in full before choosing the fix. It contains W-01 through W-30 with examples. In mixed repositories, load both only when the defect crosses those stacks.
  • For other stacks, apply the gate and seven signals directly; use the ecosystem's supported contracts and diagnostics.

The escape valve

Not every root cause is yours to fix. A workaround is allowed only when ALL hold:

1. The root cause is in external code the team does not control.
2. The proper fix needs upstream changes on an uncertain timeline.
3. The business cost of not shipping exceeds the debt incurred.
4. The workaround is isolated; it does not leak into other code.

When all four hold, contain it:

1. Mark it: // WORKAROUND: [reason]; see [issue-link or local tracking path]
2. Record an owner and removal condition locally; create an external issue only when authorized.
3. Add a test that pins the current behavior.
4. Add a focused upstream probe that detects when the workaround can be removed.
5. Set a review date (max 90 days).

If any condition fails, fix the root cause. No exceptions.

Foundations & rationalizations

When discussing the rationale or pressure to skip a root-cause fix, read references/philosophical-foundations.md. Treat its analogies as rationale, not evidence for a particular defect.

For provenance and upstream maintenance, read NOTICE.md.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.