writing@1.1.2
VERIFIED· install onlyPublisher claims: "Drafts, edits, and rewrites prose in the user's own voice — emails, posts, essays, memos, proposals, social copy. Use when asked to write, draft, edit, tighten, shorten, rewrite, or proofread something; when a draft reads flat, wordy, robotic, or "not like me"; when a piece will not start, a sentence will not come out right, or the ending trails off; when an email has to deliver bad news, say no, chase a reply, or make an ask that lands; when a post needs a hook, a headline, or a cut to fit a limit; when order, transitions, or paragraph structure are the real problem; when writing in a second language, or matching a house style, spelling variant, or word count. Learns and reuses their voice across sessions. Not for grammar-only fixes (`grammar`), sales copy (`copywriting`), book manuscripts (`book-writing`), journal papers (`academic-writing`), narrative craft (`storytelling`), versioned draft-and-audit workflows (`write`), or inbox triage and follow-ups (`email-management`)." Observed in sandbox: no network, touches nothing outside the project, writes no memory.
Registry ClawHubPublisher ivangdavila · publishing for 244 daysInstalls 9,309Last detonated 15 h ago
Coverage
Nothing executable in the bundle ran; the trace covers the install step only.- Tested
- every file read, file written and host contacted by what ran
- Not tested
- use of the skill by an agent
Is [email protected] safe to install?
KillerSkills ran it in the KillerSkills honeypot sandbox (1 scenario; nothing executable in the bundle ran; the trace covers the install step only) on 2026-11-10 and scored it 87/100 (Verified). In what ran, nothing we could detect went wrong: no undeclared network egress, no credential reads, no persistent instruction writes. What it does with a real credential was not observed; the static findings and the publisher carry the rest of the score. Summary: no network, touches nothing outside the project, writes no memory.
Behavior trace
Honeypot run · 0 s · 1 scenarios · 0 deviationst+CategoryObserved action
0.0sprocessno install scripts in bundle
0.0sprocessscript backend runs setup only
0.0sprocessscript backend runs setup only
0.0sprocessno install scripts to re-run
Static findings
LOW
Fake instruction boundary tagfeedback.md:5 · Fake instruction boundary tag · informational: not an instruction file
AST01…dback_mode` and `edit_depth` from `config.yaml`, and the relevant `style-sheets/<context>.md` if the piece belongs to a publication or client. Editing to your own taste …Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.CRITICAL
Fake instruction boundary tagSKILL.md:87 · Fake instruction boundary tag
AST01…hat number wins — check it and record it in `~/Clawic/data/writing/style-sheets/<context>.md` (`style-guide.md`). Reading time = words ÷ 240 (silent prose reading runs r…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagclarity.md:7 · Fake instruction boundary tag · informational: not an instruction file
AST01…mory.md` and the `terminology` and `banned` lines of the relevant `style-sheets/<context>.md`. A clarity edit that reintroduces a banned word is worse than no edit: it p…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Instructs the agent to modify its persistent memory or instruction filesclarity.md:135 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
AST01…ts a construction, a word, or a rewrite** — not just this once, but as a rule — write it to `## Never` in `memory.md`, and the specific edit to `## Corrections` with the trait it reveals, in the s…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Instructs the agent to modify its persistent memory or instruction filesdrafting.md:82 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
AST01….md` names it. **When a piece is commissioned, started, or given a deadline**, write its row to `## Pieces` in `memory.md` — piece, format, who for, word target, due date, status. **When a first draft …Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagsocial.md:66 · Fake instruction boundary tag · informational: not an instruction file
AST01…ks — write it to `## Formats` in `memory.md`, or to the channel's `style-sheets/<context>.md` if it is a brand or client rule. **A bio, standing intro, or reusable post …Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagsecond-language.md:29 · Fake instruction boundary tag · informational: not an instruction file
AST01… user's audience decides, and the decision goes in that context's `style-sheets/<context>.md`. ## Interference Patterns Worth Knowing Grouped by what they do to the re…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagworkplace.md:118 · Fake instruction boundary tag · informational: not an instruction file
AST01…nd is named here); a document format the organisation requires to `style-sheets/<context>.md`; and a memo, brief, handover or postmortem that will be re-read to `artifac…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagfeedback.md:63 · Fake instruction boundary tag · informational: not an instruction file
AST01… in. Record who edits and what they always change in the client's `style-sheets/<context>.md` — after two rounds it is predictable, and pre-empting it saves the round. -…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagposts.md:103 · Fake instruction boundary tag · informational: not an instruction file
AST01…ion order, banned claim, byline rule), write it to that context's `style-sheets/<context>.md` in the same turn, with the date (`memory-template.md`). Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagposts.md:71 · Fake instruction boundary tag · informational: not an instruction file
AST01- Standing furniture — a fixed opening block, a fixed closing block — is what makes an issue feel like an issue. Decide it once and write it into that newsletter's `style-sheets/<context>.md`.Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagstyle-guide.md:83 · Fake instruction boundary tag · informational: not an instruction file
AST01… **Every style decision made in a recurring context is written to `style-sheets/<context>.md` in the same turn it is made**, with its date, and the sheet's `## Boxes` li…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Instructs the agent to modify its persistent memory or instruction filessecond-language.md:72 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
AST01…inconsistency to fix. **When a recurring interference pattern is identified**, write it to `## Corrections` in `memory.md` with the trait, and after the third occurrence promote it to `## Voice` under …Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagmemory-template.md:31 · Fake instruction boundary tag · informational: not an instruction file
AST01| A style decision was made for a publication, client or project | `style-sheets/<context>.md` |Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagsecond-language.md:72 · Fake instruction boundary tag · informational: not an instruction file
AST01…irect, how formal, which greeting set — it goes to that context's `style-sheets/<context>.md` in the same turn (`memory-template.md`). Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagsocial.md:5 · Fake instruction boundary tag · informational: not an instruction file
AST01… **Before writing under a brand or a client's account**, read its `style-sheets/<context>.md` — claim boundaries, banned words, hashtag policy and who approves. Read `##…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagstructure.md:7 · Fake instruction boundary tag · informational: not an instruction file
AST01… a piece written for a publication, client or project**, read its `style-sheets/<context>.md` — required section order, length caps and mandatory blocks are structural c…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagmemory-template.md:12 · Fake instruction boundary tag · informational: not an instruction file
AST01| House style for a publication, client, employer or project | `~/Clawic/data/writing/style-sheets/<context>.md` | One file per context, from the first |Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Instructs the agent to modify its persistent memory or instruction filessocial.md:66 · Instructs the agent to modify its persistent memory or instruction files · informational: not an instruction file
AST01… that cut a post short, a limit that bit, a structure that consistently works — write it to `## Formats` in `memory.md`, or to the channel's `style-sheets/<context>.md` if it is a brand or client ru…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagvoice.md:95 · Fake instruction boundary tag · informational: not an instruction file
AST012. The style sheet for the context (`style-sheets/<context>.md`) — a client's house style outranks the user's habits inside that client's work.Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagposts.md:5 · Fake instruction boundary tag · informational: not an instruction file
AST01…riting for a publication, client, or recurring outlet**, read its `style-sheets/<context>.md` (word cap, section requirements, banned claims, byline rules) and `pieces/<…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagstructure.md:104 · Fake instruction boundary tag · informational: not an instruction file
AST01…, subhead cadence, a mandatory block — write it to that context's `style-sheets/<context>.md` in the same turn (`memory-template.md`). **When an outline outlives its pie…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagstyle-guide.md:81 · Fake instruction boundary tag · informational: not an instruction file
AST01…ghostwritten. Then record the choice as the first line of the new `style-sheets/<context>.md` — the choice of guide is itself the decision that stops being re-made. **E…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagstyle-guide.md:5 · Fake instruction boundary tag · informational: not an instruction file
AST01…riting in any recurring context**, read its `~/Clawic/data/writing/style-sheets/<context>.md`. If the context has no sheet and it will recur, the first decision made in …Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagmemory-template.md:197 · Fake instruction boundary tag · informational: not an instruction file
AST01… ## style-sheets/ One file per context at `~/Clawic/data/writing/style-sheets/<context>.md` — a publication, a client, an employer, a project, a byline. Read before wr…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagsocial.md:28 · Fake instruction boundary tag · informational: not an instruction file
AST01Character limits change. When a post is near a limit, count it rather than trusting the number here, and record any limit that surprised you in that channel's `style-sheets/<context>.md`.Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagstructure.md:102 · Fake instruction boundary tag · informational: not an instruction file
AST01- Sentence case or title case is house style, not correctness — take it from `style-sheets/<context>.md` or the `conventions` preference area.Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagposts.md:82 · Fake instruction boundary tag · informational: not an instruction file
AST01… page, write to that exact number, and record it in that outlet's `style-sheets/<context>.md`. - End on what should happen, naming who does it. An op-ed without an addre…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagvoice.md:100 · Fake instruction boundary tag · informational: not an instruction file
AST01…g with Acme's"), and a line recording the clash in that context's `style-sheets/<context>.md`. ## Voice Drift Two kinds, with different fixes: - **Their drift**: the …Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.LOW
Fake instruction boundary tagfeedback.md:66 · Fake instruction boundary tag · informational: not an instruction file
AST01…e same turn: what this reviewer or client always changes to their `style-sheets/<context>.md`, so the next draft pre-empts it; the reviewer's register and channel prefer…Fix: Remove the text that addresses the agent rather than the reader; describe behavior, do not instruct.Cost to the agent
Instructions loaded
≈6,552 tok
1 file · 15 in bundle
Tokens estimated at four characters each. Skills load their instruction text when triggered; the figure is what that costs.
Known vulnerabilities
Vulnerability advisories (OSV) cover npm and PyPI packages; this registry has no advisory database.