docs(miner): document claim ledger in README (#2291) - #1
Closed
kiannidev wants to merge 218 commits into
Closed
Conversation
…nrich hold audit metadata (JSONbored#3288) agentHoldAuditDetail's ciState==="failed" branch returned a bare, unexplained "auto-action held because CI is failing but no close action was planned" for EVERY red-CI hold, unconditionally -- it never checked protected-author or close-autonomy the way the gate-blocker-codes branch a few lines below it already did. That made a protected owner/admin/automation author and a "close autonomy isn't auto yet" repo indistinguishable from each other, and from a genuinely unexplained hold, in the single most common real-world cause of this message on a self-hosted instance. Extract the shared protected-author/close-autonomy disambiguation into closeWithheldReason() and apply it to BOTH branches, so a red-CI hold now reports "close withheld for protected author" or "close withheld because close autonomy is <level>" exactly like a gate-blocker hold already does. Also persist the previously-missing structured fields on the agent.action.hold audit event's metadata (headSha, gateBlockerTitles, ciFailingCheckNames, closeEligible, closeAutonomy, mergeAutonomy, protectedAuthor flags, closeOwnerAuthors, precisionBreakerEngaged/directions) so a hold is fully debuggable from the audit table alone. Investigated whether contributor + red CI + close=auto genuinely fails to plan a close (the reported "biggest product issue"): it does not -- planAgentMaintenanceActions' willClose condition already covers this case correctly (see the existing "blocked contributor PR ... close=auto" regression test), and this fix's own new tests confirm a hold only ever reaches the generic fallback when the author is protected AND close autonomy isn't auto -- i.e. every reachable hold now has a specific, correct explanation.
…SONbored#3296) Add high-confidence rules for bb_ (Browserbase) and ak-/as- (Modal token ID/secret) with truncation and hyphen/underscore continuation tests. Co-authored-by: Cursor <cursoragent@cursor.com>
…SONbored#3292) Extend isCodeFile and advisory isCodePath so front-end framework source is recognized consistently in slop signals, missing-tests annotations, and MCP score-preview mirrors — closing the isCodePath gap that blocked prior partial fixes. Co-authored-by: Cursor <cursoragent@cursor.com>
…ed#3294) generateReplayScoringKey never lints or scrubs frozen context, so it can only ever reject on selection -- reusing ReplayTaskRejected advertised an unscrubbable_forward_reference branch this function can never produce. Adds a regression test pinning the documented eligibility asymmetry against generateReplayTask for the same candidate.
…ps (JSONbored#3297) Fills real self-hosting documentation gaps found by audit: undocumented env vars (maintenance/installation backpressure, foreground liveness, OTEL, Orb collector overrides, per-provider base URLs, Discord/Slack PR notifications, Sentry server name, PGVECTOR_ENABLED, PORT) and .gittensory.yml blocks with zero website coverage (gate.checkMode, the CLA/expectedCiContexts/rebase-window gate fields, the anti-abuse settings cluster, contentLane, closeOwnerAuthors/blockedPaths, and the features: 3-state fallback). Adds explicit config precedence and a link to .gittensory.yml.example on the configuration page, cross-links it with the tuning page in both directions, and names SELFHOST_DEPLOYMENT_MODE explicitly for the instance-wide write switches it actually controls.
…ng (JSONbored#3291) Covers two previously undocumented self-hosting topics: how to move a repo between the hosted managed-beta App and a self-host install (what carries over via .gittensory.yml vs. what must be re-created, since settings live in separate databases), and how to cleanly decommission a self-host instance (revoke the GitHub App, choose what happens to data volumes, the current lack of a brokered-mode revocation endpoint, and removing ADMIN_GITHUB_LOGINS access).
…et-scan (JSONbored#3302) Add high-confidence rules for fal_sk_ (fal.ai) and wandb_v1_ (77-char body) with truncation and hyphen/underscore continuation regression tests. Co-authored-by: Cursor <cursoragent@cursor.com>
…JSONbored#3301) Extend isCodeFile and advisory isCodePath to recognize .cc and .hpp so native C++ changes align with review/rag.ts indexing and receive correct slop, missing-tests, and check-run annotation treatment. Co-authored-by: Cursor <cursoragent@cursor.com>
…ONbored#3306) Add high-confidence rules for xai- (xAI) and dg. (Deepgram) prefixes with truncation and hyphen/dot/underscore continuation regression tests. Co-authored-by: Cursor <cursoragent@cursor.com>
…ts (JSONbored#3303) * test(scoring): cover pure upstream-constant parsers in scoring/model.ts Closes JSONbored#2091. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(test): use literal density constants in scoring-model test Avoid Record<string, number> type errors from DEFAULT_SCORING_CONSTANTS indexing. Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com>
…JSONbored#3308) Add high-confidence rules for sk_car_/sk_car_admin_ (Cartesia) and apify_api_ (Apify) with truncation and continuation regression tests. Co-authored-by: Cursor <cursoragent@cursor.com>
…cy (JSONbored#3289) * docs(self-host): add disk capacity planning and backup retention policy The Disk section had real measured numbers from one production instance but no way to extrapolate to a larger install, and the backup-scaling doc never explained how many historical backups the `backup` profile keeps before pruning. Ground both additions in the actual growth/retention logic: review_audit and webhook_events have no retention policy and grow unbounded per PR/webhook, audit_events is capped at 90 days, and backup.sh keeps BACKUP_RETAIN (default 7) newest copies per target independently, skipping the sqlite prune entirely after a failed backup so a known-good copy is never lost. * fix(docs): tighten backup-retention and capacity-planning wording The SQLite-failure callout only named the integrity_check failure mode, missing the .backup-command-itself-fails and empty-output-file cases the script also treats as failed verification. The review_audit size estimate mixed a per-row byte figure with an aggregate MB estimate without a clear derivation; points operators at measuring their own instance instead of trusting a blanket ratio.
…#3309) Classify CMakeLists.txt, meson.build, BUILD.bazel, and MODULE.bazel as build config in path-matchers so native-build-only diffs are not miscounted as substantive source effort in slop signals. Co-authored-by: Cursor <cursoragent@cursor.com>
…on (JSONbored#3311) Closes JSONbored#2113. Co-authored-by: Cursor <cursoragent@cursor.com>
…bored#3317) Add high-confidence secret-scan rules for Baseten sky_{prefix}.{secret} federated keys (9-16 char prefix after sky_) and Laminar lmnr_ project tokens, with boundary, truncation, and continuation negative tests. Co-authored-by: Cursor <cursoragent@cursor.com>
Add high-confidence secret-scan rules for Exa exa_ search keys and Mem0 platform (m0-) and self-hosted (m0sk_) tokens, with truncation and identifier-continuation negative tests. Co-authored-by: Cursor <cursoragent@cursor.com>
…ilscale TLS profiles (JSONbored#3295) * docs(self-hosting): correct Grafana exposure claim, document Caddy/Tailscale TLS profiles security.tsx claimed Grafana's port was private by default alongside Prometheus/Qdrant/Ollama, but its compose entry publishes 3000:3000 on every interface, unlike Qdrant's confirmed 127.0.0.1-only binding -- correct the claim and note the mitigation. The caddy and tailscale compose profiles each had exactly one line of documentation anywhere; add a full TLS termination walkthrough (Caddyfile contents, DNS prerequisites, removing the app's own port mapping, bring-your-own reverse proxy, Tailscale's network_mode: host and tailnet reachability) and cross-link it from github-app.tsx's PUBLIC_API_ORIGIN/relay-mode/webhook-check sections and from quickstart's profile table. Also name and explain docker-proxy, the read-only Docker-socket proxy that isolates Promtail's service discovery from the raw socket. * fix(docs): correct Tailscale port-mapping and Caddy TLS-fallback claims Unlike Caddy, Tailscale's sidecar (network_mode: host) doesn't replace the app's listener -- it adds a new interface the existing 0.0.0.0 port-publish becomes reachable through. Telling operators to remove the ports mapping, as for Caddy, makes the app unreachable everywhere, tailnet included. Also corrects the Caddy TLS section: the shipped Caddyfile has no fallback directive, so ACME failure for a real domain does not silently substitute a self-signed cert -- only recognized non-public hostnames like localhost get that automatically. * fix(docs): scope the Caddy port-removal rule and stop implying Tailscale satisfies GitHub webhook delivery The earlier Caddy port-removal callout still named Tailscale as an alternative, directly contradicting the Tailscale section's own correction that keeping the mapping is required there. Also corrects docs.self-hosting-github-app.tsx: PUBLIC_API_ORIGIN and the wizard's webhook checks previously implied a Tailscale tailnet address could satisfy GitHub's webhook delivery requirement, which it cannot -- GitHub's servers can't reach a private tailnet address. A Tailscale- only instance needs brokered pull mode instead. * fix(docs): remove remaining no-public-port claims about Tailscale The FeatureRow summary, the section heading, its intro paragraph, the Grafana-exposure note, and the quickstart page's profile pointer all still claimed Tailscale removes public reachability on its own -- an exhaustive grep found and corrected every instance, matching the detailed explanation already added: the default port mapping keeps the app reachable on every host interface until you additionally firewall the host or use tailscale serve.
…ored#3320) Add high-confidence secret-scan rules for Braintrust bt-st- service tokens and ScrapeGraphAI sgai- UUID-shaped keys, with truncation and identifier-continuation negative tests. Co-authored-by: Cursor <cursoragent@cursor.com>
…Nbored#3321) Add high-confidence secret-scan rules for Langfuse pk-lf- public keys and Hyperbolic hb_ API tokens, with truncation and identifier-continuation negative tests. Co-authored-by: Cursor <cursoragent@cursor.com>
…ored#3290) * feat(agent): add Linear backend for project/milestone matching (JSONbored#3186) Lets a repo opt into Linear (via an encrypted per-repo API key) as the project/milestone-matching backend instead of GitHub Milestones/Projects v2. Prefers a confirmed native GitHub-link match (Linear's own attachmentsForURL) over fuzzy title/body matching, and degrades to no suggestion on any Linear API error or missing key. * test(agent): cover the linear branch of the backend selector (JSONbored#3186) Closes the codecov/patch gap flagged on PR JSONbored#3290: the "linear" arm of parseProjectMilestoneMatchBackend (repositories.ts) and the non-null arm of the settings.autoProjectMilestoneMatchBackend yml overlay (focus-manifest.ts) had no test exercising them. * test(agent): use an explicit placeholder-shaped test secret literal (JSONbored#3186) The prior fixture ("unit-test-encryption-secret-at-least-32-bytes-long") was already established convention (verbatim in ai-key-byok.test.ts on main), but the gate's deterministic generic_secret_assignment hard blocker has no path-based exemption and only skips values matching its own placeholder-keyword heuristic. Renaming to include "example" clears the pattern without changing any test behavior. * fix(agent): drop dead DB-side timestamp defaults on repository_linear_keys (JSONbored#3186) Every write to this table goes through Drizzle's $defaultFn, which always supplies the ISO timestamp explicitly, so the DEFAULT CURRENT_TIMESTAMP fallback was unreachable in practice. Removing it resolves the gate's repeated flag on this table without changing any behavior -- test/unit/linear-key.test.ts:90-98 (an insert that omits both columns) still passes unchanged.
…bored#3298) * feat(engine): honor candidatePaths in metadata lane-fit scoring When discovery metadata includes touched paths, rank metadata candidates with the same path+label goal model used in analyze instead of labels alone. Co-authored-by: Cursor <cursoragent@cursor.com> * test(engine): cover computeMetadataLaneFit branches for codecov patch Exercise non-string and blank candidatePaths through vitest so changed engine lines meet the 99% patch gate. Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com>
…#3325) Add high-confidence secret-scan rules for Gladia gla_ keys and WorkOS legacy wos_api_key_ tokens, with truncation and identifier-continuation negative tests. Co-authored-by: Cursor <cursoragent@cursor.com>
JSONbored#3327) Add high-confidence secret-scan rules for MiniMax sk-cp- subscription keys and sk-api- pay-per-token keys, with truncation and identifier- continuation negative tests. Co-authored-by: Cursor <cursoragent@cursor.com>
…#3328) * feat(engine): add pickTopRankedOpportunities ranker helper Expose a pure top-N wrapper over rankOpportunities so miners and tools can cap discovery output without reimplementing ordering or tie-breaks. Co-authored-by: Cursor <cursoragent@cursor.com> * fix(test): restore OpportunityRankInput import for typecheck The pickTopRankedOpportunities change dropped the type import while OpportunityRankInput is still used by the ranker test helpers. Co-authored-by: Cursor <cursoragent@cursor.com> * chore: retrigger CI after typecheck fix Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com>
…d#3329) categorizeFile (analysis-context.ts) recognized only .md/.mdx/.rst/.txt as docs, so the long-form spellings .markdown, .adoc, and .asciidoc fell through to the source category — a NOTES.markdown or guide.adoc change was treated as source rather than documentation. Add the three spellings to the docs extension list. This matches the canonical DOCS_EXTENSIONS set in src/signals/path-matchers.ts and rag.ts's DOC_EXT_RE, which already treat them as docs. Matching is case-insensitive via the existing lowercased-extension path (README.ADOC classifies as docs).
Rank metadata discovery candidates and return the top N entries using rankMetadataOpportunities ordering. Lives in metadata-top-pick.ts so vitest patch coverage instruments every branch. Co-authored-by: Cursor <cursoragent@cursor.com>
…ration test (JSONbored#3602) Adds a realistic fixture corpus of AI-PR-ban / near-miss docs, a fixture-driven resolveAiPolicyVerdict test over them, and a fan-out integration test proving a banned repo is hard-skipped as a pre-filter (excluded before any per-issue GitHub call, costing exactly one policy fetch). Test-only. Closes JSONbored#2306
…ed#3621) Closes JSONbored#2330 Document status/doctor walkthrough, fleet docker pattern, and invariants. Co-authored-by: Cursor <cursoragent@cursor.com>
…SONbored#2054) (JSONbored#3606) * docs(config): add minimal .gittensory.minimal.yml starter template Provide the smallest safe manifest starter with observe-only autonomy and gate off, plus self-host doc links for JSONbored#2054. Co-authored-by: Cursor <cursoragent@cursor.com> * docs(config): drop README link to avoid scope creep Keep the minimal template limited to the root manifest files. Co-authored-by: Cursor <cursoragent@cursor.com> * docs(config): add minimal .gittensory.minimal.yml starter template Provide the smallest safe manifest starter with observe-only autonomy and gate off, plus parse test and self-host doc links for JSONbored#2054. Co-authored-by: Cursor <cursoragent@cursor.com> --------- Co-authored-by: Cursor <cursoragent@cursor.com>
…d#3619) Add a local, no-network REES analyzer that flags an exported symbol a PR removes or renames in a package public entrypoint (index/mod/main/public-api) - a semver-major break for downstream consumers shipped without a major version bump. Purely diff-driven: for each changed entrypoint file it diffs the exported names on removed lines against those on added lines and reports a name that leaves the public surface, so a same-name signature/value edit is never flagged. Deterministic, conservative, and fail-safe, bounded by entrypoint and finding caps. Wires the descriptor, BriefFindings type, render section, canonical name registry, and regenerated analyzer metadata; adds unit tests covering entrypoint detection, export-name extraction, removal/rename/ edit diffing, the caps, abort, and render. Closes JSONbored#1510 Co-authored-by: e11734937-beep <e11734937-beep@users.noreply.github.com>
…review (JSONbored#3633) Adds a --profile visual-review docker-compose service (ghcr.io/browserless/chromium, internal-network-only, resource-bounded) so self-host operators can enable automated before/after PR screenshot capture without standing up their own browser-rendering sidecar. The code-level support (BROWSER_WS_ENDPOINT -> env.BROWSER injection, the puppeteer-core self-host stub) already existed; nothing was actually deployed anywhere until now. Documents BROWSERLESS_TOKEN/BROWSER_WS_ENDPOINT/etc. in .env.example, and extends scripts/smoke-selfhost.sh with an opt-in visual-review mode that boots a real browserless container and asserts /gittensory/shot returns an actual rendered PNG end to end -- verified locally (20526-byte real render of https://example.com through the full self-host chain). Entirely additive: unset/default is byte-identical to today (no container, no screenshots, no error), and docker compose config validates cleanly both with and without the profile active.
…red#1958) (JSONbored#3634) Adds a category field (security/correctness/performance/ maintainability/tests/style) to each inline finding: the model self-categorizes when asked, and a deterministic path/keyword classifier (classifyFindingCategory) supplies a safe default for whatever it omits, so a caller with the feature on never sees a partial tag. Surfaced in the inline-comment label and a new "Finding categories" collapsible in the unified comment. Gated by review.finding_categories in .gittensory.yml (default off, layered on review.inline_comments like review.suggestions) -- no new global env var, no DB/OpenAPI surface.
…SONbored#3622) Bootstrap local SQLite state with gittensory-miner init, extend doctor with sqlite and Docker checks, and document the zero-infra quickstart. Co-authored-by: Cursor <cursoragent@cursor.com>
…Nbored#3636) A narrower webhook event's embedded pull_request sub-object can omit body entirely (undefined) rather than reporting it as explicitly empty (null/""). Re-deriving linked issues from an absent body was silently wiping an already-correctly-claimed linked issue -- and resetting its claim timestamp -- on any such upsert, which could make the gate report "no linked issue" on a PR that clearly cited one in a prior, fuller sync. Falls back to the existing stored value only when body is genuinely absent; an explicit empty body still clears the claim as before.
…SONbored#3637) Add a local, deterministic deprecatedDep REES analyzer that flags a direct dependency a PR newly adds or upgrades when it is an officially deprecated or unmaintained package with a maintained successor. Reuses the shared manifest parser and matches added/upgraded names against a bundled curated list per ecosystem (npm + PyPI); no network, no token. Conservative and fail-safe: only exact list matches are reported, bounded by manifest/patch/finding caps. Closes JSONbored#1511 Co-authored-by: e11734937-beep <e11734937-beep@users.noreply.github.com>
…d#3355) (JSONbored#3638) Document that miner_claims does not adjudicate duplicate winners and add regression tests for the active-only filter (JSONbored#3354). Co-authored-by: Cursor <cursoragent@cursor.com>
…a drift check (JSONbored#3635) .gittensory.yml.example is meant to be THE exhaustive reference for every supported gate/settings/review/features/contentLane/repoDocGeneration field, but had quietly drifted: gate.aiReview.combine/onMerge/reviewers, ~10 settings fields (contributorBlacklist, linkedIssueHardRules, unlinkedIssueGuardrail, autoProjectMilestoneMatch(Backend), the 4 disposition label-name overrides, blacklistLabel, pendingClosureLabel), 12 review.* fields (footer/note/fields/enrichment/profile/security_focus/ inline_comments/suggestions/changed_files_summary/path_instructions/ instructions/pre_merge_checks/ai_model), and the entire repoDocGeneration block were never documented at all. Documents every one of them, then adds an automated exhaustiveness check (test/unit/focus-manifest.test.ts) so this can't silently drift again: each config surface gets a `satisfies Record<keyof T, string>` map from field name to its expected doc token, which is a compile-time error to leave incomplete when a new field is added to the type -- the same guarantee this repo already uses for OpenAPI/settings parity and docs-drift checks elsewhere. Also cross-links the private-config precedence doc (config/examples/README.md) to the generated SELFHOST_ENV_REFERENCE_ROWS list, since that chain only covers per-repo review policy, not deployment-level environment variables.
…bored#3351) (JSONbored#3640) Expose the foundation-phase claim ledger API names with full idempotent claimIssue coverage (JSONbored#3353). Co-authored-by: Cursor <cursoragent@cursor.com>
* fix(review): keep skill frontmatter first * fix(review): replace full generated prefix regardless of length change (JSONbored#3639) replacementStart = Math.max(0, block.startIndex - generatedStartIndex) only clamped the direction where the fresh prefix is longer than the stored one. When it's shorter (e.g. a repo rename shortens repoSkillName, or the description wording changes on a later refresh), replacementStart landed mid-string inside the old frontmatter, corrupting the file with a truncated fragment ahead of the fresh content. Whenever the renderer declares a prefix at all, the entire file up to the marker is that prefix by construction (a prefixed doc type is fully machine-generated, nothing else precedes it) -- so it's always safe to replace from byte 0 regardless of the old prefix's length. A prefix-less doc type (marker at byte 0) is unaffected and keeps preserving real content before the marker exactly as before.
… false positive (JSONbored#3641) Two independent issues surfaced by JSONbored#3635 merging concurrently with JSONbored#1958's own review.finding_categories addition: - main's typecheck is currently broken: JSONbored#3635's new compile-time exhaustiveness map (test/unit/focus-manifest.test.ts) predates review.finding_categories and is missing that key, which the `satisfies Record<...>` check requires. Documents the field in .gittensory.yml.example and adds the missing map entry, mirroring the review.changed_files_summary entry immediately above it. - "override of the synthesis merge rule" (added by JSONbored#3635) matches gittensory's own INJECTION_SOURCE pattern for "override ... the ... rule", so the review pipeline redacted it to [external-instruction-redacted] before the AI reviewer ever saw it -- correctly, since the redaction can't tell a benign config override from a real one without the surrounding context a human has. Rewording "rule" to "mode" preserves the exact meaning and clears the false positive; confirmed via a direct regex test against the full file that no other phrase matches.
Adds a history-class review-enrichment analyzer (JSONbored#1514, part of JSONbored#1499) that flags a changed file when the PR re-introduces added lines in a region a prior revert commit removed - a signal the change may be re-treading a path that was already reverted or hot-fixed out. For each changed file it reads recent commit history, detects revert commits (Revert "..." subject or a "This reverts commit <sha>" body), fetches each revert commit patch, and intersects the reverted old-file line ranges with the PR added new-file line ranges. Reports only the file, a re-introduced line, a short revert-SHA prefix, and the reverted PR number when named - never file contents. Bounded fan-out (5 files / 15 commits / 10 revert lookups / 25 findings); fail-safe on missing token, invalid slug, fetch error, or abort. Wires the descriptor into the registry, render, types, UI docs, refreshed metadata, and the analyzer-name lists, with full unit coverage. Co-authored-by: e11734937-beep <e11734937-beep@users.noreply.github.com>
* feat(review): add per-repo review.visual capture config (JSONbored#3609) Adds `review.visual` to the .gittensory.yml manifest so a self-hoster can point the before/after screenshot pipeline at their own repo's preview setup without any code changes: - preview.url_template: an explicit "after" preview URL template ({number}/{head_sha}/{head_sha_short} placeholders) that always wins over GitHub-native preview discovery -- the only option for a provider (e.g. Cloudflare Workers Builds' non-production branch builds) that never surfaces a GitHub-visible deployment at all. Validated at parse time against the same SSRF guard the renderer applies at render time (isSafeHttpUrl), so a malformed template warns immediately instead of silently degrading to a blank cell. - routes.paths / routes.max_routes: an explicit route list that replaces automatic file-to-route inference for a repo whose routing convention isn't gittensory-ui's TanStack file-based one, plus an override for the built-in per-PR route cap. All-absent (default, every repo today) resolves to the existing GitHub-native discovery chain and automatic route inference, so this is byte-identical until a maintainer opts in. Also documents the two review.* fields (finding_categories, visual) that were still missing from .gittensory.yml.example's field-exhaustiveness map after JSONbored#3635. Part of JSONbored#3607. Closes JSONbored#3609. Closes JSONbored#3610. * test(review): cover the reindented buildCapture preview-discovery branches codecov/patch flagged 91.18% on the prior commit: wrapping the existing GitHub-native discovery chain in a new `else` block (so the explicit url_template check could run first) re-indented every line inside it, which counts as "changed" for patch coverage even though the logic itself didn't change. Adds the missing branch coverage: - target.previewUrl already set (skips discovery, no fetch calls) - getLatestDeploymentStatus throwing (defense-in-depth; the callee itself always catches internally and never actually throws today) - a preview URL found via a commit check run (skips the PR-comment fallback) - every getPreviewBuildState outcome the pending-poll branch reads: building, succeeded, and absent (no matching check run at all)
…JSONbored#3649) Co-authored-by: andriypolandki <=>
…SONbored#3648) Assert column layout, NOT NULL constraints, default active status, UNIQUE (repo, issue), and CHECK rejection against the live SQLite table. Co-authored-by: Cursor <cursoragent@cursor.com>
Closes the foundation claim-ledger scaffold by documenting the local soft-claim API in the miner package README with a contract test. Co-authored-by: Cursor <cursoragent@cursor.com>
Owner
Author
|
Superseded by JSONbored#3650 |
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
16 participants
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.
Summary
openClaimLedger,claimIssue,releaseClaim,listActiveClaims) and note it is local bookkeeping only.Closes JSONbored#2291
Test plan
test/unit/miner-claim-ledger-readme.test.ts— README API surfacenpm run typechecknpm run build --workspace @jsonbored/gittensory-minerMade with Cursor