CtrlK
BlogDocsLog inGet started
Tessl Logo

writing-error-messages

Use when writing, reviewing, or rewriting user-facing error messages, validation messages, form errors, empty/error states, auth errors, failure notifications, retry/support copy, CLI errors, or API errors humans will read. Helps make errors specific, actionable, non-blaming, accessible, and safe without leaking sensitive details.

95

1.08x
Quality

92%

Does it follow best practices?

Impact

99%

1.08x

Average score across 3 eval scenarios

SecuritybySnyk

Passed

No known issues

SKILL.md
Quality
Evals
Security

Evaluation results

98%

5%

Error Message Audit and Rewrite

SaaS error message audit and rewrite

Criteria
Without context
With context

No cutesy tone

100%

100%

Specific failure description

100%

100%

Reassurance about data/state

100%

100%

Specific next action

90%

100%

Plain language over jargon

100%

100%

No blame language

100%

100%

Avoids forbidden words

75%

100%

Validation messages say how to fix

80%

80%

Empty state not treated as error

100%

100%

Backup path for persistent failures

100%

100%

Prefers 'we' over 'you' for blame

77%

100%

99%

18%

Authentication and Account Recovery Error Messages

Auth and account recovery error messages

Criteria
Without context
With context

Generic login error

100%

100%

Locked account concealment

30%

100%

Safe password reset copy

100%

100%

Expired/used reset link

100%

100%

Registration: no account enumeration

20%

100%

Field validation: how to fix

80%

100%

No blame in validation

100%

100%

Session expiry reassurance

75%

87%

Placement guidance present

100%

100%

Security notes section

100%

100%

Specific CTAs

100%

100%

100%

CLI Tool Error Messages

CLI tool developer-facing error messages

Criteria
Without context
With context

File path in config errors

100%

100%

Expected vs. received for syntax error

100%

100%

Missing field named explicitly

100%

100%

Next command or fix suggested

100%

100%

No raw exception names

100%

100%

No POSIX codes as primary message

100%

100%

Network ambiguity acknowledged

100%

100%

Permission error gives alternative

100%

100%

Typo/unknown subcommand helpful

100%

100%

Plain language first line

100%

100%

Diagnostic notes included

100%

100%

Repository
joshuadavidthomas/agent-skills
Evaluated
Agent
Claude Code
Model
Claude Sonnet 4.6

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.