avoid-ai-writing

github:conorbronsdon/avoid-ai-writing · first committed 2026-03-06 · source · badge

Audit and rewrite content to remove AI writing patterns ("AI-isms"). Use this skill when asked to "remove AI-isms," "clean up AI writing," "edit writing for AI patterns," "audit writing for AI tells," or "make this sound less like AI." Supports a detection-only mode that flags patterns without rewriting.

Skill Health: not yet scored

Listed, not yet scored. This skill has no public model test yet. It is in the scoring queue; submitted skills go first, then the most-installed.

Version history

94 commits to this file on GitHub (newest first).

2026-10-18 · commit d4be856 · Reword large-file scope sentence; document known static-scanner false positives (#410)
2026-10-15 · commit 7dc0334 · chore(release): prepare v3.38.0
2026-10-09 · commit 76a6b6e · Ground model-only audit labels in pattern catalog and fix detect-mode regressions
2026-10-04 · commit db82bc8 · chore(release): prepare v3.37.0 (#375)
2026-09-29 · commit 3454472 · Clarify incomplete residual audits and validator output compatibility
2026-09-29 · commit 307cf59 · feat: separate preservation checks from residual quality policy
2026-09-26 · commit 2a7e999 · Check rendered line breaks before delivering prose (#350)
2026-09-24 · commit b2838ed · Catch staged discoveries; extend speaker guardrail to drafting (#346)
2026-09-23 · commit 020f4a4 · docs: define language adaptation guidance and community links (#345)
2026-09-23 · commit 6135015 · chore(release): prepare v3.36.0 (#336)
2026-09-16 · commit 91b5cbe · Enforce no-op decisions and complete final scope checks
2026-09-16 · commit 0fdaf07 · Derive edit reports from the delivered final text
2026-09-16 · commit 66c997f · Merge commit '726c8c4a01a45cc126b12652a75c1c867092893c' into HEAD
2026-09-16 · commit 726c8c4 · Separate validator results from authorized scope review
2026-09-16 · commit 373f3fb · Require explicit stopping rationale and review source removals
2026-09-16 · commit 924b21c · Clarify no-op accounting and preserve source imperatives
2026-09-16 · commit 27a28dd · Make unavailable verification reporting explicit in every mode
2026-09-14 · commit dd160fb · Merge branch 'refresh/pr295' into refresh/pr296
2026-09-14 · commit b8826be · Merge commit 'fc979c6489ec0ac81a77236782ba496da81b24cb' into refresh/pr295
2026-09-14 · commit feb5268 · chore(release): prepare v3.35.0
2026-09-13 · commit a0ff473 · Return one final rewrite within a shared editing-pass budget
2026-09-13 · commit b46b1c3 · Define a source-faithful, scope-aware editing contract
2026-09-13 · commit c964d55 · docs: map context profiles to detector modes
2026-09-11 · commit e79254c · fix: prepare CLI package release
2026-09-06 · commit 66da512 · fix: treat em dashes as style-only guidance
2026-09-06 · commit 28ba61b · fix: preserve prose quote boundaries around malformed bare URLs
2026-09-05 · commit f654266 · Address release review: editable inputs, valid links, and catalog guidance
2026-09-05 · commit 0aa0def · Split canonical skill and ship complete Claude bundle
2026-09-05 · commit 3dd72f5 · Fix automatic marks inference and malformed link scanning
2026-09-05 · commit 7552fa3 · Keep explained transformations exempt from rewriting
2026-09-05 · commit b43142a · Include clarity rules in P2 triage and preserve explained summaries
2026-09-05 · commit c5ec5b8 · Complete catalog clarity rules with source-preserving guidance
2026-09-05 · commit 081b495 · Narrow repeated reversal punchlines and integrate current main
2026-09-03 · commit d8c2351 · Add detector trio from #108: launch intros, crowd contrast, fake-casual props (#128)
2026-08-29 · commit 3bd64f1 · Add seven rhetorical-tic categories from Simon Willison's LLM cliché highlighter (#138)
2026-08-28 · commit 4e79a67 · feat: add rendered Markdown source mode (#137)
2026-08-24 · commit 40328bd · docs: restrict edit mode to prose files (#136)
2026-08-24 · commit 0bdab7b · docs: document npx skills as a universal install method (#131)
2026-08-22 · commit ca2206c · Bind voice-profile targets to Never-inject guardrails (fixes #100) (#133)
2026-08-17 · commit 8277224 · Add paraprosdokian detection category (v3.26.0)
2026-08-16 · commit 6f77e68 · Add catalog-only entries from #108: false agency, transformation crutch, domain-term collision
2026-08-16 · commit 7b5fdc7 · Add detector trio from #108: launch intros, crowd contrast, fake-casual props
2026-08-13 · commit 3c0fd8a · Add actually as a cut-first writing signal (#126)
2026-08-13 · commit 46087ce · feat: detect unnecessary hyphenation (#120)
2026-08-06 · commit b72b7c4 · Write the openclaw emoji literally in SKILL.md frontmatter (#118)
2026-08-05 · commit a575602 · release: 3.23.1 — constructor false-positive fix, automated npm publishing (#117)
2026-08-04 · commit f666087 · feat: optional --style house-style layer (config-driven, points at Vale) (#99)
2026-08-03 · commit 5897f7b · skill: close five prose-contract gaps flagged in downstream review (#97)
2026-08-03 · commit ae1ecbe · detector: stop counting non-tag "#" forms as hashtags (#90) (#92)
2026-07-31 · commit 3b5bbc1 · Fix title-case-header on Markdown headings, and the FPs that fix introduced (#62)
2026-07-31 · commit fba57d3 · Enforcement, self-scan proof, and the first measured accuracy numbers (v3.22.0) (#68)
2026-07-31 · commit 94998a9 · Add narrated-candor rule (v3.21.0) (#66)
2026-07-30 · commit 3a34c68 · Add lingering-attention claims rule (v3.20.0) (#65)
2026-07-29 · commit 22529c2 · Add "verbatim" as a Tier 3 (density-gated) word (#63)
2026-07-29 · commit 4d9e851 · Add moral-adjective category errors and invented contrast-pair mirror (#61)
2026-07-22 · commit 27156c7 · v3.18.0: carve out definition-list separator em dashes (#60)
2026-07-22 · commit 3200004 · Rename "Subjectless fragments" to cover its passive-voice half; split 3.17.0 changelog entry (#59)
2026-07-22 · commit 193f4f6 · Harvest the remaining blader/humanizer patterns: 4 categories + 4 rule extensions (v3.17.0) (#58)
2026-07-16 · commit 53ae05a · v3.16.0: add "load-bearing" (metaphor) to Tier 1 (#55)
2026-07-08 · commit 500ff59 · v3.15.0: conversational-register patterns (wall-of-text, recap-flattery) (#53)
2026-07-07 · commit f072db1 · Add 'vague third-party validation' rule (#39 follow-up) (#49)
2026-07-07 · commit 04394de · Add speculative-scenario-opener detector + tropes.fyi gap-fill (#45) (#47)
2026-07-07 · commit 49a8a31 · Add "quietly" to Tier 2 word table (#44)
2026-07-05 · commit f9f8265 · Broaden the 'It's not X, it's Y' rule to catch the split-sentence form (#42)
2026-06-18 · commit 13b644c · Add "match inflected forms" rule to the word-replacement section (#37)
2026-06-18 · commit 15c1083 · Fix em-dash mojibake (U+FFFD) in skill docs (#36)
2026-06-18 · commit 98209a1 · Add "genuinely" to hollow-intensifiers rule and word table (#35)
2026-06-10 · commit 1e37593 · Add bold list-label periods detection category (#31) (#32)
2026-06-06 · commit 4331560 · Detect social endorsement closers (#29) (#30)
2026-05-29 · commit 5900248 · Add Self-labeling significance rule (v3.8.0)
2026-05-28 · commit efdbb6a · v3.7.2: calibrate curly-quotes rule per missed #15 review (#27)
2026-05-28 · commit 4190d0f · Merge main into #15 + reconcile curly-quotes rule (v3.7.1)
2026-05-28 · commit b5241a3 · v3.7.0: catalog candidates — mixed punctuation, hyphenated pairs, gap-filling, authority tropes (#24)
2026-05-28 · commit ee686b7 · v3.6.0: voice profiles + edit mode + iterate-to-convergence (#23)
2026-05-28 · commit 12b4c41 · v3.5.0: catalog adds from competitive audit (P38/P41/P43 + P40) (#20)
2026-05-27 · commit df9d18d · Add curly quotation marks rule to Formatting section
2026-05-17 · commit 3b3d663 · feat: AI-tool fingerprint sections (placeholders, citation markup, UTM) (#11)
2026-05-17 · commit a64723d · feat: position-pivot preface + Cursor rule port (#10)
2026-05-17 · commit 176b5a0 · v3.4.0: structural + crypto-shill detection patterns (#9)
2026-04-07 · commit 088c4ad · Add 'hit differently' to Tier 1 word table + emotional flatline
2026-04-01 · commit d6266f7 · Add worth-verbing and reader-steering frame patterns
2026-03-31 · commit 357be3f · Add detect mode: flag AI patterns without rewriting (v3.2.0)
2026-03-25 · commit d1b81fe · v3.1.0: Add Pangram AI detection insights — 3 words, 2 templates, structure priority
2026-03-24 · commit 9287587 · v3.0.0: severity tiers, context profiles, 4 new patterns, agentskills.io compliance
2026-03-18 · commit 9f1956f · Add OpenClaw compatibility with single SKILL.md
2026-03-18 · commit 5228f84 · Resolve merge conflicts with tiered vocabulary (2.0.0)
2026-03-18 · commit 1f5e31e · Add 7 new pattern categories, rhythm checks, and rewrite threshold
2026-03-18 · commit 1510409 · Introduce tiered vocabulary system with 102 entries
2026-03-18 · commit 049d1b2 · Add replacement table flexibility note to tone calibration
2026-03-18 · commit 20fa74b · Expand AI-ism coverage: new words, dedup fillers, flag "let's" pattern
2026-03-18 · commit f6a5dc9 · Detect double-hyphen (--) as em dash substitute
2026-03-06 · commit 6d0f0e4 · Add pattern #22: emotional flatline
2026-03-06 · commit db1b742 · Add 8 pattern categories, original full example, per-pattern before/after tables
2026-03-06 · commit ec2f036 · Initial release: AI writing audit skill for Claude Code

Archived snapshots