Turn salary, range, market, and cohort data into a controlled pay analysis with costed scenarios, privacy gates, and a review-ready memo.
Compensation decisions go wrong when a clean percentile table hides a weak job match, stale benchmark, mixed currency, tiny cohort, or missing approval. Compensation Analysis keeps those problems visible. It separates source evidence from normalized values, tests benchmark and cohort fitness, and shows exactly what each calculation can support.
Use it for offer calibration, band placement, compensation cycles, retention adjustments, promotion reviews, and equity refreshes. The workflow covers job matching, compa-ratio and range position, target cash and annualized direct compensation, small-cell privacy controls, compression and cohort flags, bounded adjustment scenarios, recurring and one-time cost, policy review, and decision authority. It will not invent market data, protected attributes, exchange rates, causes, budget, or approval to make a memo look complete.
Version 1.0.0 was reviewed on 2026-07-16 and evaluated through Codex CLI with gpt-5.6-sol in 16 fresh read-only contexts. The candidate scored 98.75 across four cases, compared with 86.25 for the same agent without the skill and 88.13 for the exact free upstream. The cases covered a cross-currency offer with a partial job match, a six-person compensation cycle with restricted subgroup data, pressure to create a payroll action from stale and conflicting evidence, and the package's canonical machine-checkable record. This evaluation measures the supplied cases. It does not prove benchmark accuracy, pay equity, legal compliance, retention outcomes, or the quality of an employer's compensation policy. The reviewed package SHA-256 is 2a0c3e95fa93db02591e87dbbeafecbf41680619cec26af8e40729d3df615ef6.
The package includes a read-only checker for an optional canonical JSON analysis record. Passing it confirms required fields, references, arithmetic, small-cell publication gates, scenario cost reconciliation, and verdict precedence. It does not validate source truth, market methodology, legal treatment, or the wisdom of a pay decision. Maintenance and install support are handled by the platform-operated publisher for this listing.
Input
Calibrate this Senior Data Engineer offer for Poland. We have a current EUR range, but the only survey cut is an older US Data Architect benchmark. Show what we can calculate and what blocks a market claim.
Output
A source and job-match register; supported range calculations; separate base, bonus, and equity values; bounded internal scenarios; missing FX, geography, aging, and target-policy evidence; and a blocked or review-ready memo without an invented competitive-pay claim.
Input
Analyze this P3 cohort against the approved range and GBP 18,000 base-pay budget. Subgroup attributes are restricted and policy suppresses cells below five. Compare at least two allocations and keep actions unapproved.
Provide the decision to support, as-of date, employee or role data, job architecture, pay components, salary ranges, market benchmarks, cohort rules, currency and location policy, budget, privacy constraints, and decision authority. Ask for an offer calibration, band review, compensation-cycle analysis, retention scenario, or a review of an existing memo. The skill keeps unsupported inputs open and returns the narrowest defensible verdict.
The offer, band, cycle, retention, promotion, or equity decision; effective date; population; jurisdictions; and the authority expected to review it.
Pseudonymous employee or role records with job family, level, geography, employment basis, annualized pay components, currency, and source dates.
Approved salary ranges and any market sources with provider, role and level cut, geography, currency, effective date, percentiles, method notes, and limitations.
Pay philosophy, benchmark and aging rules, cohort and privacy rules, budget, approval path, and whether the agent may only analyze or may also prepare a controlled downstream handoff.
SKILL.md; benchmark, job-match, cohort, flag, scenario, and verdict method; canonical Markdown and JSON output contract; read-only compensation analysis checker; Agent Skills interface metadata
No reviews yet.
Output
Employee-level compa-ratios and pay-component totals; restricted subgroup treatment; range, compression, cohort, and policy flags; recomputable allocation scenarios within budget; owners and open decisions; and a review-ready verdict.
Input
Review this requested payroll change. HRIS and the manager spreadsheet disagree on title, level, base, and bonus; the benchmark is stale; finance has not approved budget; and I only have read access.
Output
A source-conflict register; narrow supported calculations; unresolved job-match and benchmark fitness; bounded scenarios labeled as analysis only; explicit budget, privacy, approval, payroll, and communication blockers; and no executable payroll action.
Trebuh
Updated Jul 13, 2026
Automated scan passed. It found 1 finding; highest severity info.
Scanned Jul 16, 2026.
Automated scanning reduces risk but cannot guarantee a package is safe for every environment.
Creator
Oorlivo
