summarize-pro@2.0.7
VERIFIED· install onlyPublisher claims: "A faithful summarizer that picks one of five output format families (minutes / comparison / timeline / bullets / cards) by reader intent, then compresses under a strict no-hallucination discipline: nothing absent from the source, ambiguity flagged rather than resolved, every figure copied verbatim. Ships a zero-dependency, offline toolchain (extract / citecheck / selfcheck / redact / validate / normdates / neutral / render) plus an 22-case golden set, turning 'we do not fabricate' from a claim into a runnable assertion: citecheck traces each output sentence back to source tokens and lists anything untraceable, selfcheck asserts format-family completeness including mandatory strength labels, extract builds the skeleton and additionally grades decisions and commitments - every conclusion is labelled firm / provisional / proposal / conditional and every action item is labelled committed / named_unconfirmed / intent_only, so 'what was actually decided' and 'who actually took it' become checkable instead of assumed. Conditional clauses are kept out of the action table and markdown headings out of the sentence stream. Standard library only, no network, no replacement for human judgement. Also ships **deterministic rendering** (receipt / personal / executive / bundle views, top-3 per section), **relative-date anchoring** that appends an ISO date without rewriting the source sentence, and a **neutral-language check** for personnel / medical / legal / incident scenes. Fidelity is checkable: citecheck traces numbers to source, and the golden set re-runs with an exit code. Triggers: summarize, summary, meeting minutes, minutes of meeting, decision log, action items, comparison table, timeline, key points, digest. Use when turning a transcript, meeting recording, interview or long document into minutes, a comparison table, a timeline or a brief." Observed in sandbox: no network, touches nothing outside the project, writes no memory.
Coverage
Bundled scripts ran with no arguments; no agent task scenario for this bundle.- Tested
- 2 of 2 bundled scripts run with no arguments, with decoy credentials in the environment · every file read, file written and host contacted by what ran
- Not tested
- use of the skill by an agent
What changed since 2.0.6
vet-skills diff summarize-pro 2.0.6 2.0.7In the sandbox, this version reached for nothing 2.0.6 did not. Compared from the two recorded runs, not from source.
The guard can hold an update like this until someone approves it (runtime.on_change: hold). A change that only shows with a real credential, or on a path the run did not reach, is not in a trace and so not in this comparison.
Is [email protected] safe to install?
KillerSkills ran it in the KillerSkills honeypot sandbox (2 scenarios; bundled scripts ran with no arguments; no agent task scenario for this bundle) on 2026-11-18 and scored it 93/100 (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 · 2 scenarios · 0 deviationsStatic findings
No findings across 2,833 lines in 22 files.
Cost to the agent
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.