dtsong/agentic-council · Archived

finance-tax-research

Use when answering a federal, state, or international tax question that requires authoritative research. Covers IRC, Treasury Regs, Rev Rul/Proc, case law, and applicable treaty positions. Do not use for tax-provision computation (use finance-provision — Phase 3+).

First seen Jun 12, 2026

Installation

$ npx skills add dtsong/agentic-council --skill finance-tax-research

Stronger alternatives

This repository is archived — consider an actively maintained alternative.

Similar popular skills

Related neighbors and high-traction skills in the same topics — useful to compare before installing.

Also in this package

Other skills from dtsong/agentic-council.

npx skills add dtsong/agentic-council

Browse all from dtsong/agentic-council

More details

Agent compatibility

Declared targets from SKILL.md / docs. Unmarked agents are not listed — the skill may still install via the CLI.

Claude Code Not declared
Cursor Not declared
Codex Not declared
GitHub Copilot Not declared
Windsurf Not declared
Gemini CLI Not declared
Cline Not declared
OpenCode Not declared

Repository health

License LICENSE
Default branch main
Open issues 0
Status Archived

Skill metadata

Parsed from SKILL.md frontmatter.

Version1

Package contents

Files included with this skill beyond the listing page.

  • skill md SKILL.md 7,830 B
  • docs SUMMARY.md 295 B

History

  1. First seen on skills.sh
  2. First recorded snapshot · 1 installs

SKILL.md

Finance Tax Research

Purpose

Answer a federal, state, or international tax question with a defensible authority-based memo. The deliverable cites the controlling authorities in hierarchy order, applies them to the facts, reaches a tentative conclusion, and flags any uncertain-tax-position implications under FIN 48 / ASC 740-10.

Scope Constraints

  • Produces a research memo and UTP recommendation; does not compute the tax provision (handoff to finance-provision once available) and does not file a return.
  • Covers federal, state, and international (treaty-based) research; does not cover transfer-pricing benchmarking studies (specialized work).
  • Does not constitute legal advice; the memo is a working paper for the tax function and outside counsel review.

Inputs

  • The factual question, framed by the requester (with proposed transaction, dates, parties, amounts)
  • Applicable jurisdictions (federal / state / foreign)
  • Taxpayer entity classification (C-corp, S-corp, partnership, foreign branch, etc.)
  • Relevant prior-year returns or positions for consistency
  • Any open IRS / state examinations or letter rulings on point
  • Authoritative research database access (BNA, CCH, RIA, Westlaw, Lexis)
  • Materiality / risk threshold for UTP evaluation (per firm policy)

Input Sanitization

All inputs are read-only artifacts. No user-supplied values are interpolated into commands or file paths. Citations are verified against the authoritative source (not paraphrased from secondary sources) before being included in the memo.

Procedure

Progress Checklist

<!-- Track completion across compaction boundaries -->

  • Step 1: Frame the Issue as a Yes/No or Quantitative Question
  • Step 2: Identify Controlling Authority Hierarchy
  • Step 3: Pull Authorities and Quote Relevant Text
  • Step 4: Apply Each Authority to the Facts
  • Step 5: Identify UTP Implications under FIN 48 / ASC 740-10
  • Step 6: Draft Research Memo

Step 1: Frame the Issue as a Yes/No or Quantitative Question

Restate the requester's question in research-memo form: a precise yes/no question ("Is the proposed transaction eligible for IRC §351 nonrecognition?") or a quantitative question ("What portion of the §163(j) limitation applies to interest on intercompany debt?"). Vague questions yield vague memos; pin the issue down before pulling authority.

Step 2: Identify Controlling Authority Hierarchy

Map the issue to the authority hierarchy:

  1. Statute — Internal Revenue Code (federal), state code, treaty text
  2. Regulations — Treasury Regulations (final > temporary > proposed)
  3. Revenue Rulings — IRS published interpretive guidance
  4. Revenue Procedures — IRS procedural guidance
  5. Case law — Tax Court, district court, circuit court, Supreme Court (note jurisdiction)
  6. Private Letter Rulings / TAMs — non-precedential but instructive
  7. Secondary sources — BNA portfolios, CCH explanations, treatises (citable for context, not authority)

Note where authorities conflict. Higher-tier authority controls.

Step 3: Pull Authorities and Quote Relevant Text

For each authority that bears on the issue:

  • Full cite (e.g., IRC §351(a); Treas. Reg. §1.351-1(a)(1); Rev. Rul. 2003-51, 2003-1 C.B. 938)
  • Quote the operative text verbatim (not paraphrased)
  • Note the date adopted or modified, and any subsequent guidance

Step 4: Apply Each Authority to the Facts

For each authority, apply it to the taxpayer's facts:

  • What the authority requires / permits / prohibits
  • How the facts satisfy or fail each prong
  • Where the facts are ambiguous, identify the assumption and flag it

Reach a tentative conclusion only after all bearing authorities have been applied.

Step 5: Identify UTP Implications under FIN 48 / ASC 740-10

If the conclusion involves an uncertain position:

  • Apply the two-step FIN 48 test: (a) more-likely-than-not (>50%) recognition threshold, (b) measurement at the largest amount > 50% likely of being realized.
  • If uncertain → recommend UTP reserve and quantify
  • Flag whether the position requires Schedule UTP disclosure (taxpayer size threshold)

Step 6: Draft Research Memo

Memo structure: Issue / Conclusion / Authorities / Analysis / UTP Implication.

Memo is the workpaper of record; archive with full cite list and quoted authority text.

Compaction resilience: If context was lost during a long session, re-read the Inputs section to reconstruct the question and jurisdiction, check the Progress Checklist for completed steps, then resume from the earliest incomplete step.

Handoff

  • If the conclusion drives a current- or deferred-tax computation, hand off to the future finance-provision skill for ASC 740 workpaper preparation.
  • If the position triggers a UTP reserve, hand off to the Auditor specialist (and finance-controls-audit) for ICFR implications around tax controls.
  • If the position implicates SEC disclosure (e.g., material UTP reserve, change in deferred-tax valuation allowance), hand off to RegRep / finance-disclosure-language for footnote and MD&A language.

Output Format

# Tax Research Memo: [Short Title]

## Header
- **Requester:** [name / function]
- **Date:** [date]
- **Jurisdiction(s):** [federal / state / foreign]
- **Taxpayer:** [entity, classification]
- **Researcher:** [name]
- **Reviewer:** [name]

## Issue
[Precise yes/no or quantitative question]

## Conclusion
[1-2 sentences. Direct answer to the Issue. Note confidence level — "more likely than not," "should," "would," "will" — per the firm's opinion-language conventions.]

## Authorities
| # | Cite | Type | Operative Text (quoted) |
|---|------|------|-------------------------|
| 1 | [IRC §...] | Statute | "..." |
| 2 | [Treas. Reg. §...] | Regulation | "..." |
| 3 | [Rev. Rul. ...] | Rev Rul | "..." |
| 4 | [Case, jurisdiction] | Case law | "..." |
| 5 | [PLR / TAM] | PLR (non-precedential) | "..." |

## Analysis
[Paragraph(s) applying each authority to the facts. Identify any factual assumptions and their impact on the conclusion.]

## UTP Implication (FIN 48 / ASC 740-10)
- **Position uncertain?** [yes / no]
- **MLTN recognition threshold met?** [yes / no — with reasoning]
- **Measurement (if recognized):** [largest amount > 50% likely of being realized]
- **Reserve recommendation:** [amount or N/A]
- **Schedule UTP disclosure required?** [yes / no — with size-threshold note]

## Follow-Up to Provision Team
- [Specific computation, deferred-tax impact, valuation-allowance implication, or disclosure follow-up]

Quality Checks

  • Issue is a precise yes/no or quantitative question, not a topic
  • Authority hierarchy followed (statute > reg > Rev Rul > case > PLR > secondary)
  • Operative authority text is quoted verbatim and cite-verified
  • Conclusion uses calibrated opinion language ("more likely than not" / "should" / "would" / "will")
  • Each material authority is explicitly applied to the facts (not just listed)
  • FIN 48 / ASC 740-10 two-step test addressed for any uncertain position
  • Schedule UTP disclosure threshold considered
  • Researcher ≠ reviewer per SoD policy

Evolution Notes

<!-- Observations appended after each use -->