progressive-passes

name	progressive-passes
description	Full 3-pass progressive deep-reading of a paper — skim, grasp, targeted verify — producing a structured bundle ready for fact-checking and article drafting. This is the default reading tactic for the paper-reading package; use it whenever no angle is already decided and you're starting from a bare paper reference.
execution	tactic
dependencies	{"sops":["first-pass-skim","second-pass-grasp","third-pass-verify","extract-structured-bundle"]}

Progressive Passes

Keshav's 3-pass method, adapted into 3 sequential sops plus a finalization step:

1. first-pass-skim — 5-minute skim, classify paper type, draft candidate angles
2. second-pass-grasp — deep-read core sections, draft the bundle
3. third-pass-verify — targeted re-read for any bundle field still uncertain after pass 2
4. extract-structured-bundle — finalize into the typed schema

Available SOPs

• first-pass-skim (subagent) — pass 1
• second-pass-grasp (subagent) — pass 2
• third-pass-verify (subagent) — pass 3, only for fields flagged uncertain
• extract-structured-bundle (subagent) — finalization

Execution Guidance

• Run all 4 sops in sequence — this tactic has no branching of its own (v1 has only this one deep-read tactic; targeted-extraction is deferred, see plan Deferred Scope).
• third-pass-verify may be a no-op if second-pass-grasp didn't flag anything uncertain — that's expected and not an error.

Available SOPs

Optional, no fixed order; the final leaf is always a sop.

SOP	When to use
first-pass-skim	Five-minute skim pass to classify paper type and draft candidate angles.
second-pass-grasp	Deep-read core sections to draft the bundle.
third-pass-verify	Targeted re-read for uncertain bundle fields.
extract-structured-bundle	Finalize bundle into the typed schema.