Session value investigation showed adoption (2.7k invocations, 10+
repos) but zero findings→fix conversions; the fix loop is now the
focus.
Tasks
Notes
Released v0.9.0 (10-09): .gitignore-aware walk by
default (pretender-094 — respect_gitignore=true + --scan-ignored; kills
the ./-anchoring exclude-list footgun specodelic hit), executable
release tarballs + curl install docs (PR #42), genesis-vibes 0.12.3
(pretender-ipn, zero code changes). crates.io + GH release + binaries
live; tap/scoop both GREEN for the first time (TAP_GITHUB_TOKEN fix
validated, pretender-2tp closed).
Released v0.8.0 (10-01): managed-block provenance +
drift doctor check (genesis 0.11.1), tree-sitter-rust 0.24 (&raw
silent metric-loss fix), format-aware --staged short-circuit. crates.io
+ GH release at 0.8.0; tap/scoop patched manually (5585fbf / 5a5227a —
token still missing).
Released v0.7.1 (10-01): advisory-lease fail-closed
diagnostics (GH #35).
Released v0.6.0 (09-29): resolution tracking,
pre-push hooks, hook-location doctor check, genesis 0.8.3. crates.io +
GH release + tap/scoop all at 0.6.0 (tap/scoop patched manually —
workflow token missing, see pretender-2tp).
Notes
Anti-goal on rule work: thresholds/config must stay byte-identical;
noise drops must come from parser/rule-scope fixes only (verified via
git diff).
Rule semantic changes require an openspec proposal first (project
constraint).
pi-session evidence base for adoption metrics:
~/.pi/agent/sessions/ (2,701 pretender check
occurrences across 156 session files as of 09-20).