chore(pre-launch-review): pre-launch documentation-hygiene pass - #64
Merged
Conversation
…tragglers
Finding 2 (external review) — verify overclaim/proof language is fully retired.
Verification:
grep -rniE "mathematical(ly)? prov|formally verif|totality proof|provably" \
--include="*.md" --include="*.rs" --include="*.py" . --exclude-dir=.git
Two hits, both "provably" (no "mathematical proof"/"formally verified"/
"totality proof" anywhere):
- THREAT_MODEL.md:463 "garbage in, *provably* garbage out" (a limitation
disclosure, but proof-adjacent) -> "garbage in, garbage out".
- ts_cli/src/main.rs:5187 rustdoc "capabilities the daemon provably never
needs after BPF load" -> "... never needs after BPF load".
Neither described the Z3 pipeline as a proof; both are worded away from proof
language to match the established constraint-check register. Post-fix grep: 0 hits.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…DME perf
Finding 3 (external review). The latency section's plain-English gloss ended
with an editorializing comparison: "faster than the blink of an eye (~100 ms)
by ~400x".
Verification: `grep -n "blink of an eye" README.md` -> README.md:63.
Removed only the blink comparison; kept the plain-English unit translation
("about a quarter of a millisecond"), which serves non-technical readers, and
left the P50/P95/P99 latency table and its measured numbers untouched. The
other two "In plain English" glosses (throughput, attack-suite) translate units
without editorializing and were not changed.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Finding 4 (external review). LAUNCH_CHECKLIST.md was tracked in the public repo
root and publicly visible on gh/main. Verification:
git ls-files --error-unmatch LAUNCH_CHECKLIST.md -> tracked
git cat-file -e gh/main:LAUNCH_CHECKLIST.md -> present on public main
Its "Manual" section is distribution strategy, not engineering hygiene:
L40 native-LinkedIn video posting mechanics (repo URL in first comment)
L42-43 "Warm DM to a named contact for a repost into the AFWERX/DIU-adjacent
network"
That is a go-to-market playbook and does not belong in the public tree.
Action:
- git rm LAUNCH_CHECKLIST.md
- .gitignore it (plus LAUNCH_OUTREACH.local.md, the local-notes copy)
This content should live in the PRIVATE fleet repo (jinn-guard-enterprise) or
local notes, not the public repo.
Audit-trail note: the file also held the CVE-2026 -> JG-ADV-* rename record
(Finding 1). That trail is preserved independently in SECURITY/ADVISORIES.md
(JG-ADV-2026-001..004 with dates, status, and fix commits), which the README
already cites as the canonical registry, so removing this file loses no advisory
history. Only the historical duplicate-003 renumber footnote was unique to this
file; ADVISORIES.md is canonical going forward.
History note: `git rm` stops tracking but the file REMAINS in git history and on
the public remote. Scrubbing history (git filter-repo) and/or rotating anything
sensitive is a separate, outward-facing operator decision — not done here, not
pushed.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…independent audit"
Finding 5 (external review) — "independent audit" wording collision. RR-001 in
RESIDUAL_RISKS.md is authoritative: NO independent third-party security audit
exists yet. Two tracked docs mischaracterized the commissioned June 2026
claim-calibration review as "independent".
Verification:
grep -rniE "independent(ly)?[^.]{0,60}(audit|review|calibrat)" \
--include="*.md" --include="*.rs" --include="*.py" . --exclude-dir=.git
Adjudicated:
- Offenders (fixed):
ACTION_MANIFEST.md:22 "the independent calibration audit's" -> "the
commissioned claim-calibration review's"
ACTION_MANIFEST.md:25 "the one net-new capability the audit identified"
-> "... the review identified"
CHANGELOG.md:57 "an independent review note" -> "an external
claim-calibration review note"
- Correct, kept verbatim: every line stating an independent audit does NOT
exist / is the open gate (RR-001, THREAT_MODEL.md:11/194/522, README.md:653,
BENCHMARKS-02.md:127, OWASP-MAPPING.md:141, PROFESSOR_VALIDATION.md,
COVER_NOTE.md:40, CHANGELOG.md:76/428).
- Legit "independent" (left): README.md:634 "Independent reviewers:",
scripts/validate/* "independently verify", provenance_manifest.rs:622
"independent re-walk".
- Untracked/internal (not launch-facing, left): COMPLIANCE.md,
LICENSING_AND_BRANDING_PLAN.md, JINN_GUARD_ARTICLE_WHITE_PAPER.md,
JINN_GUARD_TEST_AND_PUBLIC_STATUS_REPORT.md (line 111 is an "Avoid:" list),
analysis/*.
RR-001 unchanged. Post-fix: no tracked text implies a third-party security audit
occurred.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
…rtifacts
Finding 6 (external review) — additive publication, no deletions.
Verification: the six review artifacts existed on disk under
analysis/analysis-report/ (untracked) and were referenced by ACTION_MANIFEST.md
via that untracked path. `git ls-files` confirmed none were tracked.
Action:
- Copied verbatim (byte-identical, cmp-verified) into
docs/calibration-review-2026-06/: EXECUTIVE_MEMO, CLAIM_EVIDENCE_MATRIX,
GOVERNANCE_MATURITY_SCORECARD, C2PA_READINESS_GAP_MAP, TECHNICAL_APPENDIX,
VALIDATION_STATUS. Findings, scores, and "Overstated" labels are UNMODIFIED.
- Added docs/calibration-review-2026-06/README.md: states this was a
commissioned claim-calibration review (June 2026), NOT a third-party security
audit (defers to RESIDUAL_RISKS.md RR-001), and carries a gap->closure table.
Table cites only real ancestor hashes: proof/"semantic firewall" wording +
plane-split guarantees table 9ff64f1 (PR #50 1b0b22b), claim-hygiene cff00da,
guarantees un-stale 134dc68 (PR #60 0696529), Action Manifest v0 94be991
(PR #52 f9c889b) + key-pin c2af362. Three C2PA gaps with no closing commit
(external transparency anchoring; hardware-backed signer identity; full
machine-readable claim taxonomy) are marked Open, not invented.
- Repointed ACTION_MANIFEST.md's C2PA gap-map reference from the untracked
analysis/analysis-report/ path to docs/calibration-review-2026-06/.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
tattoosonmyskin
force-pushed
the
chore/pre-launch-review
branch
from
July 16, 2026 01:52
6c8e8ca to
ee15835
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Pre-launch documentation-hygiene pass. Verify -> report -> fix only what's confirmed,
one commit per finding with the verification evidence in each commit message. Docs-only
(the single
.rschange is a rustdoc comment);cargo check --workspacepasses.Findings
provablyproof-language stragglersTHREAT_MODEL.md+ts_cli/src/main.rsrustdoc reworded off proof language; post-fix grep clean6a9c8ea7dc4b26LAUNCH_CHECKLIST.md(LinkedIn tactics, warm-DM to AFWERX/DIU-adjacent contact)git rm+.gitignore; belongs in the private fleet repo / local notes. History still retains it (filter-repo is a separate operator decision)3cd2c47ACTION_MANIFEST.md+CHANGELOG.mdreworded to "commissioned/external claim-calibration review"; RR-001 unchanged; no tracked text now implies a third-party security auditdbf884cdocs/calibration-review-2026-06/(findings/scores/"Overstated" labels unmodified); added an index README that states this was a commissioned review not a third-party security audit (defers to RR-001) with a gap->closure table citing only real ancestor hashes and 3 C2PA gaps marked Open; repointedACTION_MANIFEST.mdto the new in-repo path6c8e8caVerified clean, no change needed
LAUNCH_CHECKLIST.mddoc reference (only the intentional.gitignoreentry).Notes for the operator
LAUNCH_CHECKLIST.mdremains in git history and on the remote; scrubbing (git filter-repo) is a separate outward-facing decision.mainstill carries two earlier ad-hoc commits that overlap this branch's README/checklist fixes; recommendgit branch -f main gh/mainso this PR is the single source, or just let the merge supersede them.🤖 Generated with Claude Code