What changed
HYPOTHESIS (not supported by the provided signals): enterprises are reportedly evaluating coding agents for Java framework migrations (ScarfBench, per the convergence description) while open-source intent-tracked agent git workflows (Aura) create a substrate for intent metadata. CRITICAL DATA-QUALITY NOTE: the two source signals actually attached to this convergence — an IRS estate-tax closing-letter fee rule and an HTMLAreaElement spec fix — are unrelated to the claimed causal chain. The load-bearing signals (80, 161) were referenced but their sources were not provided, so every premise here is unverified.
Why now
IF agents begin executing high-liability migrations at regulated enterprises, change-management frameworks (SOC2 CC8, ITGC change controls) would require documented rationale for each change, and agent-authored commits lack a human author to interview. This is a plausible second-order obligation, but it is an inferred future need — no auditor guidance, regulation, or customer complaint in the input demonstrates it exists today.
Converging signals
Claimed: agent viability for enterprise migrations + open-source intent-metadata git workflows. Verified from provided sources: none — the attached signals (IRS estate tax fee, HTMLAreaElement hreflang) do not support the convergence. Graph cluster c3 (cohesion 0.917) suggests semantic relatedness in the corpus, but that is engine-internal, not market evidence.
Customer pain
HYPOTHESIS: compliance leads cannot produce a defensible change-rationale trail when an agent authored hundreds of migration commits. demand_evidence is EMPTY — zero complaints, zero job postings, zero mandates were retrieved. Per the system's own lesson (capability-rich, demand-blind engine), absence here is weak evidence of absence, but the rule stands: demand must be scored from evidence, and there is none.
Who pays
Claimed buyer: engineering managers and compliance leads at regulated enterprises adopting agent-assisted migrations. This is precisely the buyer class the founder profile excludes as a sole channel — regulated-enterprise compliance tooling is bought through security review, vendor onboarding, and procurement, not through demonstrated-value self-serve.
Solved today
HYPOTHESIS: today, teams either prohibit agent-authored changes in regulated codebases, require human review/sign-off on every agent commit (making the human the auditable author), or rely on PR descriptions and ticket links as the intent trail. Notably, 'human reviews and merges the PR' already satisfies most SOC2 change-management controls, which undercuts the new-obligation premise.
Why current solutions are bad
Manual review of bulk agent migrations doesn't scale and the intent-to-commit linkage is lossy. But this is inconvenience, not a compliance gap an auditor has flagged — no evidence in the input shows an audit finding, failed SOC2 control, or budget line attached to this problem.
Proposed product
A CLI/CI tool + hosted report generator that ingests agent intent metadata (Aura-style records, PR descriptions, task tickets) and git history, cryptographically links each commit to its originating business intent, and emits a signed, auditor-readable migration provenance report. Per-migration-report or per-seat pricing.
MVP version
Open-source CLI that parses an intent-metadata format plus git log and renders a static HTML/PDF provenance report for one migration; a paid hosted tier adds signing, retention, and SOC2-mapped control language. 3-5 weeks of solo AI-assisted build.
30-day build
Validate before building: interview 10 platform/compliance leads at companies piloting coding agents; search SOC2 auditor guidance and Big-4 publications for agent-change-control language; instrument demand (GitHub issues on Aura/agent repos asking for audit trails). Build nothing beyond a landing page and sample report until one buyer says 'we failed or fear an audit over this.'
60-day build
If (and only if) validation finds a named buyer with an audit deadline: ship the CLI + one hosted report flow, integrate with GitHub PRs and one agent framework, and get 2 design partners using it on a real migration.
90-day revenue plan
Charge design partners $500-1500 per migration report package. Realistic first-revenue is 6-9 months given enterprise security review — outside the founder's 180-day window unless a mid-market design partner short-circuits procurement.
Distribution path
Weak: open-source CLI + content on 'auditing agent-authored code' + integration listings (GitHub Marketplace). No forced buyer, no existing watering hole of compliance leads reachable without relationship sales.
Pricing hypothesis
$500-1500 per migration report or $200-500/mo per team hypothesized; zero pricing evidence in input.
Technical difficulty
Low-moderate for the founder: git plumbing, metadata parsing, report generation are squarely in AI-assisted solo range. The hard part is not technical — it's trust and standardization (whose intent format wins).
Legal / regulatory risk
Low direct legal risk, but the product's value depends on auditors accepting its reports — an acceptance the founder cannot manufacture and has no fire-service/industrial-credibility analog for in SOC2 audit circles.
Platform dependency
High: depends on agent frameworks emitting intent metadata (Aura's format is new open source, could change or be absorbed) and on GitHub. GitHub, Cursor, or Anthropic adding native audit trails would erase the wedge — audit logging is a natural first-party feature.
Founder fit
Poor. This is not the proven government-portal-mandate shape: no regulation compels anyone to file anything, there is no portal, no per-filing transaction, and no deadline. The buyer is enterprise compliance — a long-trust-cycle, procurement-gated channel the founder explicitly avoids. His ELDT edge (forced filers + federal portal + per-upload fee) shares only the word 'compliance' with this idea.
Breakout potential
If agent-authored code provenance becomes a formal audit requirement (plausible in 1-3 years), the category could be real — but incumbents (GitHub, GitLab, agent vendors, Vanta/Drata) are far better positioned to own it as a feature.
Final recommendation
KILL for now; revisit if concrete evidence appears (an auditor or framework formally requiring agent-change provenance, or hiring/spend data showing teams paying for it). The idea is coherent and solo-buildable, but it fails on the right kill criteria: no demonstrated buyer, no urgent pain, incumbent-absorbable wedge, and an enterprise-procurement-only channel that mismatches the founder. This is not a capital problem — runway cannot buy a demand signal that doesn't exist yet.
Next action
Do not build. Add a watch: track SOC2/ISO auditor publications, Aura adoption, and job postings mentioning 'AI change management' or 'agent audit trail'; re-score when demand_evidence is non-empty.