Environmental Policy Researcher

Environmental Policy Researcher

When to Use

Use this skill when:

• User wants to research a specific bill, ordinance, or regulation by name or number (e.g., "What does the Clean Air Act Section 111(d) actually require?" or "How does SB 100 work in California?")
• User is preparing to submit formal public comment on a proposed rule, environmental impact statement, or regulatory docket
• User wants to understand the mechanics of a specific policy instrument -- how cap-and-trade systems set price floors, how a renewable portfolio standard calculates compliance, or how a Clean Water Act Section 401 certification works
• User needs to evaluate competing claims about a policy's projected costs, benefits, or feasibility using primary evidence
• User is preparing to vote on a ballot measure with environmental components and wants to understand what it actually does versus what advocates and opponents claim
• User wants to map the stakeholder landscape for a policy debate -- who is involved, what they argue, what evidence they use, and where their interests lie
• User is a student, journalist, or engaged citizen who needs to build a rigorous, evidence-based understanding of an environmental policy question without being told what to think about it
• User is confused by contradictory information in news coverage or advocacy materials and wants help sorting factual claims from framing and spin

Do NOT use when:

• User is asking which environmental policies they personally should support -- this skill builds research capacity, not advocacy positions; stop before recommending positions
• User needs climate science explained (physics of greenhouse gases, attribution studies, IPCC conclusions) -- use climate-science-explainer instead
• User wants practical sustainability actions for their home or lifestyle (energy audits, diet changes, transportation choices) -- use home-energy-audit, carbon-footprint-estimator, or related skills
• User's business must comply with a specific environmental permit, regulation, or enforcement action -- this is legal and regulatory compliance work, not citizen policy research
• User needs help designing a lobbying campaign, writing political fundraising materials, or organizing a ballot initiative drive -- this skill explicitly avoids partisan strategy
• User is asking about international treaty frameworks (Paris Agreement, Kyoto Protocol, Convention on Biological Diversity) as diplomatic instruments -- those require a distinct foreign policy and international law framework
• User wants to understand the underlying science behind an environmental problem (how particulate matter damages lungs, how nutrient loading causes hypoxia) without reference to policy -- science education is a separate task

Process

Step 1: Clarify the Research Question

Before doing any research, establish what the user actually needs to know and why. These questions shape everything:

• Policy level and jurisdiction: Is this a federal rule (EPA, Army Corps, USFWS), a state statute or regulation, a regional compact (RGGI, Western Climate Initiative), a county ordinance, or a city zoning code amendment? Each level has different legislative processes, different official sources, and different timelines.
• Stage of the policy: Is this a bill (not yet law), a proposed rule (in notice-and-comment period), an enacted statute (passed but possibly being implemented), a final rule (published in the Code of Federal Regulations or state equivalent), or a policy under litigation or review?
• Specificity of user's question: Does the user have a bill number, agency docket number, or policy name? Or are they exploring a topic area (e.g., "how do states regulate agricultural runoff")?
• User's purpose: Voting on a ballot measure requires different depth than a graduate seminar paper; writing a formal public comment requires knowledge of the legal standard for that comment process; preparing for a city council meeting requires knowing who already has input on the record.
• User's baseline knowledge: Policy-naive users need the mechanism explained before they can evaluate evidence. Users who follow policy actively may need only targeted deep dives.

Clarify all five dimensions before proceeding. If the user has a specific bill or rule, ask for the number or sponsoring agency. A bill number like H.R. 5376 or a docket number like EPA-HQ-OAR-2021-0317 anchors all subsequent research in primary documents.

Step 2: Identify the Policy Mechanism

Environmental policies use a small number of identifiable mechanisms. Naming the mechanism correctly is not academic -- it predicts who bears costs, who benefits, how compliance works, and what the enforcement points are. Apply this taxonomy precisely:

Command-and-Control (CAC) Regulation

• Government sets a specific performance standard, technology standard, or outright prohibition
• Non-compliance triggers penalties (fines, permit revocation, injunctions, criminal liability)
• Examples: NAAQS (National Ambient Air Quality Standards) set concentration limits for six criteria pollutants; CAFE standards set fleet-average fuel economy thresholds; RCRA sets waste handling requirements
• Key questions: What is the exact standard? Who must comply (source category, size threshold, SIC code)? What is the attainment deadline? What are the penalty tiers?

Market-Based Instruments

• Cap-and-trade: government sets a total emissions cap, issues tradable permits, lets the market determine where reductions happen. Price of a permit (allowance) is a live market signal. The RGGI cap covers CO2 from power plants in participating northeastern states; California's cap-and-trade covers ~85% of state GHG emissions.
• Carbon tax / fee: government sets a price per ton of CO2-equivalent; entities pay the fee or reduce emissions. British Columbia's carbon tax started at CAD $10/tonne in 2008, escalated to $65/tonne by 2023. The policy creates certainty on price but not on the quantity of emissions reduced.
• Tradable permits (other pollutants): the Clean Air Act's Acid Rain Program used SO2 allowances; the EPA's Cross-State Air Pollution Rule uses similar instruments for NOx.
• Key questions: What is the cap trajectory? What is the current allowance price? Are there price floors/ceilings (cost containment reserves)? What offsets are permitted?

Subsidies, Tax Incentives, and Direct Investment

• Investment tax credits (ITCs): the federal ITC for solar installations was 26% of system cost in 2022, extending to 30% through 2032 under the Inflation Reduction Act for residential and commercial solar
• Production tax credits (PTCs): wind energy PTCs are calculated per kilowatt-hour of electricity generated, incentivizing ongoing production rather than upfront installation
• Grants and loan guarantees: DOE Title XVII loan guarantee program for clean energy projects; USDA REAP grants for rural renewable energy
• Conservation payments: USDA Conservation Reserve Program pays farmers per acre per year to take highly erodible land out of production
• Key questions: Who is eligible? Is the incentive refundable or only an offset against tax liability? What is the budget cap or expiration date? What activities are excluded?

Performance Standards vs. Technology Standards

• Performance standards specify an outcome (emissions rate must not exceed X grams per kilowatt-hour) and allow any technology to achieve it
• Technology standards specify a required technology (best available control technology, maximum achievable control technology) -- more prescriptive, less flexible
• New Source Performance Standards (NSPS) under CAA Section 111 are performance standards. BAT requirements under CWA effluent guidelines are typically technology-based.

Information and Disclosure Requirements

• Toxic Release Inventory (TRI): facilities above threshold quantities of listed chemicals must report annual releases publicly; the data is searchable by facility and chemical
• Energy performance certificates for buildings (ENERGY STAR scores for commercial buildings; home energy scores)
• SEC climate risk disclosure rules require large public companies to report material climate-related financial risks
• Key questions: What must be disclosed? To whom? On what schedule? Is the information publicly searchable?

Voluntary Programs

• ENERGY STAR: EPA-run certification for appliances, buildings, commercial equipment; no mandate, but widely adopted
• Green power purchasing programs; EPA Green Power Partnership
• Farm Bill conservation programs with voluntary enrollment (EQIP, CSP)
• Key questions: What do participants actually commit to? What verification occurs? What are participation rates in the relevant sector?

Liability and Remediation

• CERCLA (Superfund): joint and several liability for contamination cleanup; PRPs (potentially responsible parties) include current and past owners, operators, generators, transporters
• RCRA corrective action: facilities with hazardous waste permits must clean up past contamination
• Natural resource damage claims: federal and state trustees can sue for damages to public natural resources
• Key questions: Who is liable? What cleanup standard applies (risk-based vs. background)? What does the remediation cost, and who actually pays when PRPs are insolvent?

Step 3: Map the Legislative and Regulatory Process

Understanding where a policy is in its lifecycle determines what research is timely and what actions are possible.

Federal Legislation:

• Bill introduction → committee referral → committee markup (often where bills are substantially amended) → floor vote in one chamber → same process in the other chamber → conference (if versions differ) → presidential signature or veto
• Key primary source: Congress.gov -- search by bill number, keyword, or sponsor; see full text, all amendments, committee reports, CBO scores, and votes
• CBO (Congressional Budget Office) scores estimate 10-year fiscal impact of legislation using specific scoring conventions; understand that CBO scores do not capture macroeconomic effects under static scoring unless the committee requests dynamic scoring
• Committee reports (H. Rpt. or S. Rpt.) explain congressional intent and are used by courts and agencies to interpret ambiguous statutory language

Federal Rulemaking (the Administrative Procedure Act process):

• Agencies have authority to regulate only within the bounds of their statutory authorization (enabling statute)
• Proposed Rule: published in the Federal Register with a docket number (e.g., EPA-HQ-OAR-2022-0829); public comment period typically 30-60 days, sometimes extended
• Comments submitted by regulated industries, environmental groups, state agencies, and individuals are part of the official record and must be responded to in the Final Rule's preamble
• Final Rule: published in the Federal Register, codified in the Code of Federal Regulations (40 CFR for EPA, 30 CFR for MSHA, 43 CFR for BLM, etc.)
• Final rules are subject to Congressional Review Act disapproval (rarely used but important when control of Congress changes); to judicial review in federal courts
• Primary source: regulations.gov -- search by docket number; read proposed rule preambles, which explain the agency's legal authority, technical rationale, and response to anticipated comments

State Legislation:

• Most states have searchable legislative databases (legistar systems for city councils, official legislature websites for state bills)
• State bills go through similar committee processes; bicameral states (all except Nebraska) require passage in both chambers
• State agencies promulgate regulations under state administrative procedure acts; state registers and administrative codes are the equivalents of the Federal Register and CFR
• State environmental agencies (e.g., CARB in California, NYSDEC, TCEQ in Texas) often have more stringent or more industry-friendly rules than federal minimums

Local Ordinances:

• City councils and county boards adopt ordinances; often searchable on Legistar platforms or official municipal websites
• Environmental components often appear in zoning codes, building codes, stormwater management ordinances, or local solid waste regulations
• Local planning commission hearings and environmental review processes (CEQA in California, SEQRA in New York, or NEPA for federally funded local projects) generate substantial public record

Step 4: Identify and Evaluate Sources

Not all sources are equal, and in environmental policy the same underlying data can support contradictory narratives depending on framing. Teach the user to evaluate source type, methodology, and potential conflict before accepting any claim.

Primary Sources (highest evidentiary value):

• Actual bill or rule text (not a summary of it)
• Official agency technical support documents and regulatory impact analyses (RIAs)
• CBO cost estimates and EPA economic analyses
• Peer-reviewed scientific literature (not press releases about studies)
• Government data repositories: EPA's AirData, EPA's ECHO (enforcement and compliance data), USGS National Water Information System, EIA energy data, NOAA climate data

Secondary Sources (useful but require evaluation):

• Congressional Research Service (CRS) reports: prepared by nonpartisan CRS analysts, consistently high quality, cover both science and policy. Often linked from Congress.gov bill pages.
• GAO (Government Accountability Office) reports: independent audits and evaluations of federal program implementation; rigorous methodology, clearly documented scope
• National Academies of Sciences, Engineering, and Medicine reports: consensus documents by expert panels; no policy recommendations, but strong on scientific and technical findings
• University-based policy centers (Resources for the Future, Nicholas Institute at Duke, Haas Energy Institute at UC Berkeley): publish peer-reviewed and working paper policy analyses; note that some receive industry funding

Advocacy Sources (useful for understanding positions, require framing analysis):

• Environmental Defense Fund, Sierra Club, NRDC, Earthjustice: publish policy analyses reflecting environmental values; often technically rigorous but advocacy-framed
• American Petroleum Institute, National Association of Manufacturers, U.S. Chamber of Commerce: publish cost analyses and regulatory impact critiques; often technically detailed but reflect member interests
• Rule: check who funded the study. The American Legislative Exchange Council (ALEC) drafts model legislation for states; knowing a bill originated from ALEC's template is relevant context. The same applies to model policies from environmental organizations.

Red Flags in Policy Research Sources:

• A study that presents only costs but not benefits (or vice versa)
• Projections that use worst-case inputs throughout without sensitivity analysis
• Comparison to other jurisdictions without controlling for context (e.g., comparing energy prices in states with very different fuel mixes)
• Reliance on a single outlier study when multiple studies reach different conclusions
• Executive summaries that misrepresent the nuance in the body of the same document

Quality Checklist for Any Policy Claim:

1. Is the underlying data identified and accessible?
2. Does the methodology handle uncertainty with ranges, not single-point estimates?
3. Has the analysis been subjected to peer review or at minimum independent technical review?
4. Does the organization's funding create a plausible interest in the conclusion?
5. Do other independent analyses reach similar conclusions?
6. Is the claim about empirical effects (jobs lost, tons of pollution reduced) or about values (how much a clean river is "worth")?

Step 5: Conduct the Stakeholder Analysis

Every significant environmental policy debate has an identifiable stakeholder map. Mapping it completely prevents the error of thinking there are only two sides.

Stakeholder Categories:

• Regulated entities: Companies, industries, or individuals who must comply with the policy. Subdivide by size: large corporations often support regulations their compliance infrastructure can handle but that create barriers for smaller competitors; small businesses often face proportionally higher compliance costs. Note whether regulated entities are concentrated (a handful of major facilities) or diffuse (millions of small sources).
• Implementing agencies: Federal EPA, state environmental agencies, Army Corps of Engineers, USFWS, BLM. Agencies have institutional interests in their authority, budget, and administrative workload. An EPA that lacks enforcement staff will implement differently than a well-resourced state agency.
• Environmental advocacy organizations: Range from mainstream (Environmental Defense Fund favors market-based instruments) to grassroots (local air quality coalitions often prefer direct regulation). Do not treat "the environmental community" as monolithic -- EDF and the Center for Biological Diversity frequently disagree on policy design.
• Environmental justice communities: Communities disproportionately exposed to pollution -- often low-income, often communities of color -- have distinct interests that may differ from mainstream environmental groups. They may support stricter regulation but oppose trading programs that could concentrate pollution near their communities (the "hot spot" problem in cap-and-trade).
• Labor unions: Interests vary sharply by industry. Unions representing fossil fuel workers (UMWA, USW) may oppose climate regulations; unions representing building trades may support clean energy incentives that create construction jobs.
• State and local governments: Often regulated entities themselves (municipal water systems, transit agencies) while also serving as regulators and enforcers. States with different economic profiles often divide into regional coalitions in federal policy debates.
• Scientific and academic community: Generally not a unified voice, but consensuses emerge through meta-analyses, National Academies reports, and IPCC-style assessment processes. Note when a scientific claim has a clear consensus vs. when it is an area of active research with genuine uncertainty.
• Consumers and the general public: Represented through polling data, public comment records, and advocacy group membership. Public opinion polling on environmental issues shows persistent majority support for environmental protection across partisan lines, but support weakens when cost impacts are emphasized.

For each stakeholder, document:

• Stated position (support, oppose, support with modifications, neutral pending specifics)
• Primary concern (economic cost, competitive fairness, health impact, community effect, job impact, property values, scientific validity)
• Evidence they cite (specific studies, industry analyses, economic models)
• Organizational capacity (a Fortune 500 company's detailed regulatory comment carries more weight in legal proceedings than a form-letter campaign, but both are in the record)
• Potential conflicts of interest (financial, ideological, political)
• History with this policy area (did they support or oppose predecessor policies?)

Step 6: Evaluate the Evidence for Effectiveness

Effectiveness claims in environmental policy require specific analytical scrutiny. Apply these tests:

Counterfactual problem: Every policy claim about effects requires comparison to what would have happened without the policy. "Emissions fell 20% after the policy passed" is meaningless without knowing what the baseline trend was. Look for studies that use difference-in-differences methodology, synthetic control groups, or other causal identification strategies -- not just before/after comparisons.

Cost estimates: Regulatory impact analyses (RIAs) are required for major federal rules under Executive Order 12866. They must quantify costs and benefits in dollar terms where possible. Check:

• Are costs measured as total social cost or just compliance cost to regulated entities?
• Do cost estimates account for technological learning (costs often fall significantly as industries adopt new technologies at scale)?
• Are job loss estimates net (accounting for jobs created) or gross?
• What discount rate is used for future costs and benefits? (Discount rates matter enormously for long-lived environmental policies; a 7% rate treats a benefit 30 years from now as worth only 13 cents compared to a dollar today; a 3% rate values it at 41 cents.)

Benefit estimates: Benefit quantification in environmental policy uses several methodologies:

• Avoided mortality: the EPA values a statistical life (VSL) at approximately $11.6 million (2022 dollars); this figure is controversial but grounded in labor market and contingent valuation studies
• Hedonic pricing: estimates environmental value from property price differentials
• Travel cost method: estimates recreational value from what people spend to visit natural areas
• Contingent valuation: surveys asking willingness to pay for environmental improvements; valid methodology but sensitive to framing

Precedent analysis: When a policy has been implemented elsewhere, real-world data beats projections. When evaluating precedents:

• Control for jurisdiction characteristics (a California climate policy can't be evaluated by comparing California's economy to Mississippi's -- many other factors differ)
• Identify the time elapsed since implementation -- some effects (particularly economic adjustment effects) take 5-10 years to fully manifest
• Note whether the comparison jurisdiction implemented the policy faithfully or had significant carve-outs, delays, or enforcement gaps

Distributional analysis: Who benefits and who bears costs is as important as the aggregate numbers. A policy that produces $10 billion in aggregate economic benefits but concentrates harms in a single low-income community is different from one with the same aggregate effects distributed broadly. Look for distributional analysis in RIAs; note when it is absent.

Step 7: Assess Implementation Feasibility

A well-designed policy can fail at implementation. Assess feasibility concretely:

Regulatory authority: Does the implementing agency have clear statutory authorization? Major federal environmental rules are routinely challenged on the grounds that the agency exceeded its statutory authority. The Supreme Court's 2022 decision in West Virginia v. EPA limited EPA's authority to require grid-wide "generation shifting" under CAA Section 111(d). Know what legal challenges a rule faces or is likely to face.

Agency resources: Enforcement capacity matters. EPA has approximately 14,000 employees and a budget of roughly $9 billion (FY2023) -- significant, but spread across air, water, land, toxics, and enforcement programs for 330 million people. State agencies implement most federal environmental programs under delegation agreements; check whether the state has accepted delegation and whether it has adequately funded the program.

Monitoring and enforcement: How will compliance be measured?

• Continuous emissions monitoring (CEMs) provide real-time stack data for major point sources -- high data quality, expensive to install
• Periodic stack tests: less continuous but acceptable for smaller sources
• Ambient air/water monitoring: measures environmental outcomes but doesn't directly attribute to specific sources
• Self-reporting with auditing: most common; accuracy depends on penalties for false reporting

Implementation timeline: From statute to enforceable standard, the federal rulemaking process typically takes 3-7 years for complex rules. State implementation plans (SIPs) required under NAAQS can take additional years. Factor this timeline into assessments of when a policy would actually take effect.

Interaction effects with existing policy: New policies layer onto existing regulatory frameworks. A state carbon price might interact with federal fuel economy standards, utility rate regulation, and renewable portfolio standards -- with complex effects that single-policy analyses miss.

Step 8: Help the User Synthesize and Position

The final step is helping the user organize the evidence into a coherent picture that supports their purpose (voting, commenting, understanding, writing). Be explicit about the structure of the synthesis:

Distinguishing scientific consensus from policy debate:

• The climate system is warming due to human GHG emissions: scientific consensus, not legitimately disputed in this skill
• How fast to reduce emissions and at what cost: policy question with legitimate disagreement
• Whether a specific cap-and-trade design will achieve projected emissions reductions: empirical claim, evaluate with evidence from comparable systems
• Whether the economic disruption costs are "worth it": value judgment, reflects priorities, cannot be answered with evidence alone

Identifying the load-bearing claims: In any policy debate, some factual claims are central to the entire argument, and others are peripheral. Help the user identify which claims, if false, would undermine a major stakeholder's position. This focuses research on high-leverage questions.

Naming value trade-offs explicitly: Environmental policy regularly involves:

• Present costs vs. future benefits (infrastructure investment, climate mitigation)
• Concentrated costs vs. diffuse benefits (a steel plant closing vs. regional air quality improvement)
• Certainty vs. flexibility (technology mandates provide certainty but lock in current solutions; performance standards allow innovation but don't guarantee specific technologies)
• Aggregate efficiency vs. distributional equity (cap-and-trade may achieve emissions reductions at lower aggregate cost but may concentrate pollution in disadvantaged communities)

These trade-offs are not resolvable with evidence -- they reflect what the user values. Make that explicit rather than pretending there is a "correct" answer.

Output Format

## Environmental Policy Research: [Policy Name or Topic]

### Research Question
[The precise question being investigated -- specific enough to be answerable]

### Policy Snapshot
| Parameter                | Value                                                   |
|--------------------------|--------------------------------------------------------|
| Policy area              | [Energy / Air quality / Water / Land use / Waste / Wildlife / Climate] |
| Government level         | [Federal / State: name / Regional compact / Local: jurisdiction] |
| Policy stage             | [Pre-introduction / Introduced / In committee / Passed one chamber / Enacted / Proposed rule / Final rule / Under litigation] |
| Primary mechanism        | [Command-and-control / Cap-and-trade / Carbon tax / Subsidy / Disclosure / Voluntary / Liability] |
| Enabling statute (if rule)| [Statute name and section]                             |
| Official identifier      | [Bill number, docket number, CFR citation, or ordinance number] |
| Key deadline or vote date | [Date and significance]                               |

### How the Policy Works (Mechanism Detail)
[Plain-language explanation of the specific regulatory mechanism: what it requires, of whom, 
by when, with what penalties or incentives. Include the specific standard or threshold where known.
Reference the actual statutory or regulatory text section.]

### Stakeholder Map
| Stakeholder              | Position          | Primary Concern               | Key Evidence Cited                  | Conflict of Interest     |
|--------------------------|-------------------|-------------------------------|-------------------------------------|--------------------------|
| [Name stakeholder precisely] | [Support/Oppose/Conditional] | [Their specific argument] | [Studies or data they cite] | [Financial/political interest] |
| [Stakeholder 2]          |                   |                               |                                     |                          |
| [Environmental justice communities] | | | | |

### Evidence Assessment
| Claim                    | Evidence Source     | Methodology           | Strength          | Limitations              |
|--------------------------|---------------------|-----------------------|-------------------|--------------------------|
| [Specific factual claim] | [Primary source ID] | [Study design type]   | [Strong/Moderate/Weak/Contested] | [What the evidence can't tell us] |

### Cost-Benefit Analysis
| Factor                   | Estimated Magnitude | Methodology Used   | Who Bears It / Who Benefits | Timeframe     | Confidence Level |
|--------------------------|---------------------|--------------------|-----------------------------|---------------|------------------|
| Cost: [Specify precisely] | [$ range or %]     | [How calculated]   | [Which group]               | [Short/long term] | [High/Medium/Low] |
| Benefit: [Specify precisely] | [$ range or quantified units] | [VSL, hedonic, etc.] | [Which group] | [Timeframe] | [Confidence] |
| Distributional note      | [Who gains, who loses disproportionately] | | | | |

### Implementation Assessment
| Factor                   | Assessment                                                |
|--------------------------|-----------------------------------------------------------|
| Implementing agency      | [Agency name, relevant division or program office]        |
| Statutory authority      | [Legal basis; note if contested]                          |
| Compliance monitoring    | [How compliance is measured -- CEMs, stack tests, self-reporting] |
| Enforcement mechanism    | [Civil penalties, permit revocation, criminal liability; penalty tiers] |
| Timeline to implementation | [Rulemaking timeline, attainment dates, phase-in schedule] |
| Interaction with existing rules | [How this layers on or replaces existing requirements] |
| Legal vulnerability      | [Pending or likely litigation challenges]                 |
| Precedent jurisdictions  | [Where similar policies exist; age of policy; measured results] |

### Scientific Consensus vs. Policy Debate
| Question                 | Type                     | Current Status                            |
|--------------------------|--------------------------|-------------------------------------------|
| [Is X a problem?]        | Scientific/empirical     | [Consensus / Active research / Disputed]  |
| [Will policy Y fix X?]   | Empirical policy claim   | [Evidence from comparable cases]          |
| [Is cost Z acceptable?]  | Value judgment           | [Legitimately contested -- depends on priorities] |

### Research Gaps and Open Questions
1. [Specific factual question that would change your evaluation if answered differently]
2. [Data that would be needed to resolve a contested empirical claim]
3. [Aspect of the policy design that is unclear from available sources]

### Forming Your Position
- **What the evidence clearly supports:** [Factual claims with strong evidentiary basis]
- **Where evidence is mixed or uncertain:** [Contested empirical questions]
- **Value trade-offs that only you can weigh:** [Where priorities, not evidence, determine the answer]
- **The strongest argument for support:** [Steelman the pro-policy case]
- **The strongest argument against:** [Steelman the anti-policy case]
- **Next research steps for this specific question:** [What to look up, where, and why]

Rules

1. Never state or imply a policy position. This skill builds research capacity. The AI's role ends at "here is what the evidence shows and here are the trade-offs." The moment the AI says "you should support this" or "this policy is clearly the right approach," it has violated the skill's core purpose. Phrase conclusions as "the evidence is consistent with," not "this proves we should."

2. Always name the specific policy mechanism before evaluating the policy. Cap-and-trade and a carbon tax are both market-based instruments but have fundamentally different political economy, price certainty profiles, and distributional effects. Treating them as equivalent because both "put a price on carbon" produces confused analysis.

3. Separate scientific consensus claims from policy claims at every stage. The physics of CO2 forcing, the toxicology of lead, and the epidemiology of particulate matter health effects are scientific consensus claims -- the AI should not artificially balance them with industry-funded contrarian views. The question of what regulatory response is appropriate is a policy claim where genuine disagreement is legitimate.

4. Primary sources over summaries, always. A CBO score should be cited from the CBO's actual report, not from a party press release summarizing it. A study's conclusions should be verified in the abstract or paper, not accepted from an advocacy summary that may misrepresent them. Press releases about research are not the research.

5. Cost-benefit analysis must always be symmetric. It is intellectually dishonest to present only compliance costs for regulated industry without also presenting quantified environmental and health benefits, and vice versa. If the user has seen a source that presents only one side, call that framing out explicitly without dismissing the source entirely.

6. Discount rates and time horizons must be stated explicitly. A $100 benefit 30 years from now is worth approximately $41 today at a 3% discount rate and approximately $13 at a 7% discount rate. Most cost-benefit disagreements in long-horizon environmental policy reduce, partly, to disagreements about the appropriate discount rate. Always identify what discount rate a study used.

7. Environmental justice analysis is mandatory for any policy affecting pollution distribution. Even if the user doesn't ask about it, a complete policy analysis must address whether the policy could concentrate pollution in communities already bearing disproportionate environmental burdens. This is not an optional add-on -- it is a required dimension of any complete analysis.

8. Implementation feasibility is as important as policy design quality. A well-designed policy that cannot be monitored, enforced, or resourced will not achieve its goals. Always assess the agency's capacity, the monitoring methodology's reliability, and the penalty structure's deterrent effect -- not just the stated goals of the policy.

9. Precedent jurisdictions must be evaluated in context. When citing another jurisdiction's results (e.g., "Sweden's carbon tax reduced emissions"), identify how similar the jurisdictions are in terms of economic structure, energy mix, population density, and policy context. Do not use a cherry-picked success or failure story as a definitive predictor.

10. Acknowledge genuine scientific uncertainty without manufacturing it. Some policy debates involve genuine scientific uncertainty (the precise ecosystem effects of a specific pesticide at low doses; the exact climate sensitivity range). Others involve manufactured uncertainty campaigns designed to delay policy action. Distinguish them by looking at the weight of peer-reviewed evidence, the consensus position of relevant professional scientific societies, and whether "dissenting" scientists have relevant credentials and peer-reviewed publications in the field.

Edge Cases

The user wants help writing a formal public comment on a proposed federal rule. This is one of the most impactful civic actions an individual or organization can take, and it has specific legal requirements. Help the user understand:

• Comments must be submitted to the official docket (regulations.gov) before the comment deadline; late comments may not be considered
• A legally effective comment is specific: it identifies the proposed rule by docket number, cites the specific regulatory provision being addressed, and presents evidence or legal argument rather than mere expressions of opinion
• Agencies must respond to all "significant" comments -- a technically substantive comment backed by data or legal argument is far more likely to influence the final rule than a form letter
• The administrative record of a rule (all comments, agency responses, technical support documents) becomes evidence in subsequent court challenges; a comment that documents data gaps or methodological flaws can support later litigation
• Help the user identify: What specific provision do they want to comment on? What evidence or argument supports their position? What data does the agency appear to have overlooked?

The user is evaluating a ballot measure and the official voter guide pro/con arguments seem misleading. Ballot measure voter guides in most states require official pro and con arguments, but these are written by campaign proponents and opponents, not by neutral analysts. Help the user find:

• The official fiscal impact analysis (required in most states; prepared by the nonpartisan legislative analyst, Department of Finance, or equivalent)
• The full text of the measure (not the title, which is often written by one side)
• Independent academic or policy institute analyses (many state universities publish ballot measure analyses before elections)
• Who is funding the campaigns for and against (required disclosures in campaign finance filings; searchable through state campaign finance databases)
• The teach the user that ballot measure titles and summary language are often written by the sponsoring side and can be misleading -- always read the actual operative text

The user encounters a peer-reviewed study that seems to contradict scientific consensus. This is common -- outlier studies exist in every scientific field. Help the user contextualize it:

• Is the study published in a peer-reviewed journal? (Peer review is a floor, not a ceiling -- predatory journals exist; check impact factor and journal reputation)
• Have other researchers replicated or failed to replicate the finding?
• Is the study an outlier in its field? What does the weight of the literature say?
• Are the study authors' affiliations or funding sources disclosed? Did a regulated industry fund a study finding no harm from its product?
• If the study is genuinely new and high quality, it represents scientific debate within the research community -- note it as such, not as a reason to dismiss the consensus, but as a reason to track the literature

The user is researching a policy that has been caught in prolonged litigation. Major environmental rules are routinely challenged in federal court and may be stayed (blocked from taking effect) while litigation proceeds. Help the user:

• Identify which court has jurisdiction and what stage the case is at (preliminary injunction, merits briefing, circuit court appeal, cert petition)
• Understand the standard of judicial review for agency rules: courts review agency statutory interpretation under the Chevron framework (now modified by Loper Bright Enterprises v. Raimondo, 2024, which ended Chevron deference) and review agency factual findings under "arbitrary and capricious" review
• Note that a rule under litigation may not be in effect, partially in effect (some provisions stayed while others are not), or effective with modifications ordered by a court
• Help the user find the relevant court docket (PACER for federal courts; most have public docket searches) to understand the case status

The user is researching a proposed policy at the local level (city council, county commission) where the public record is sparse. Local policy often lacks the structured documentation that federal rulemaking provi

…(truncated)