From e7871450bedbbdd5e522d8480edd0a95b7e37577 Mon Sep 17 00:00:00 2001 From: Sam Catlow Date: Wed, 2 Sep 2026 23:40:29 +1000 Subject: [PATCH] chore: sync issue tracker entries and metadata --- .beads/issues.jsonl | 6 ++++++ .beads/metadata.json | 3 ++- 2 files changed, 8 insertions(+), 1 deletion(-) diff --git a/.beads/issues.jsonl b/.beads/issues.jsonl index cea0d458d..ca539691b 100644 --- a/.beads/issues.jsonl +++ b/.beads/issues.jsonl @@ -91,6 +91,12 @@ {"_type":"issue","id":"archie-core-g7tm","title":"telegram: finalize the reply through a live context, matching the error branch","description":"The success path calls live.finalize(turnCtx, reply) with the raw cancellable turn context (telegram.go:582), while the error branch two lines above deliberately uses context.WithoutCancel(turnCtx). If /stop lands during finalize, the rich edit, the plain-text retry, and the split-send fallback all abort with context.Canceled (go-telegram/bot builds requests with NewRequestWithContext), so the authoritative reply is never delivered and the live message keeps the ▌ cursor forever. The reply is persisted to the ledger, so the loss is delivery-level — but the user sees a frozen half-answer.\n\nFailure: a long tool-using turn is /stop-ped as it completes; turnCtx is cancelled mid-finalize; every send fails on the dead context; the chat shows the last live frame with a blinking cursor and no final answer.\n\nFound by code-review of PR #476 (feat/467, merged 98e8dda).","design":"live.finalize(context.WithoutCancel(turnCtx), reply) — the one-liner the error branch already uses. Consider making finalize take a live context internally rather than trusting each caller to remember.","acceptance_criteria":"A /stop landing mid-finalize still delivers the authoritative reply and drops the cursor; a test covers the cancellable-context finalize path.","status":"closed","priority":1,"issue_type":"bug","assignee":"Sam Catlow","owner":"sam@catlow.tech","created_at":"2026-08-14T14:43:08Z","created_by":"Sam Catlow","updated_at":"2026-08-14T15:07:36Z","started_at":"2026-08-14T14:53:28Z","closed_at":"2026-08-14T15:03:28Z","close_reason":"Fixed in internal/channels/telegram/live.go and telegram.go: finalize/abandon strip context cancellation before sending (mirroring each other), ToolCall escapes Markdown metacharacters via bot.EscapeMarkdown, splitLongMessage splits on rune boundaries. All new/changed tests verified red-then-green; task check clean; adversarial review pass found no blocking issues.","external_ref":"https://github.com/samcharles93/archie-core/issues/484","labels":["code-review","streaming","telegram"],"dependency_count":0,"dependent_count":0,"comment_count":0} {"_type":"issue","id":"archie-core-tb3a","title":"Fix streaming review blockers for inline tool calls","description":"Resolve the four evidence-gated findings in merged inline tool-call streaming: preserve the authoritative Telegram reply when final edits fail; keep Telegram network I/O off the model FullStream consumer and cancellable; correlate tool inputs with results and render a bounded redacted parameter summary; snapshot show_tool_calls per live reply to avoid reload races. Add regression tests for every path and run task check plus lint/race validation.","acceptance_criteria":"Regression tests fail on the merged implementation and pass after the fix. Final edit failure falls back to normal split delivery. Delta and ToolCall callbacks do not wait on Telegram HTTP and stop cancels in-flight rendering. Completed tool events include bounded redacted parameters correlated by tool-call ID. Reload cannot race an in-flight reply's show_tool_calls setting. task check passes, golangci-lint reports no new findings, and focused race tests pass.","notes":"Implemented in worktree /tmp/archie-core-tb3a on branch fix/archie-core-tb3a. Added cancellable async Telegram rendering, final edit failure fallback through normal split delivery, call-ID-correlated bounded conservative parameter summaries, atomic per-reply show_tool_calls snapshots, SSE/UI parameter rendering, rebuilt ui/dist, and regression tests. Validation: GOFLAGS=-buildvcs=false task check passed; golangci-lint --new-from-rev=HEAD ./... reported 0 issues; focused race suite passed; fresh adversarial reviewer reported PASS with zero findings. Plain task check is blocked in linked worktrees by Go VCS stamping exit 128, hence buildvcs override.\n2026-08-24 reconciliation: the close reason's 'worktree left uncommitted' is stale. The work did land on main via ba33527 (Merge pull request #477 from samcharles93/fix/archie-core-tb3a); verified on main: gateway/stream.go sensitiveParameterKey+redactToolParameters, telegram/live.go per-reply showToolCalls snapshot, webui/api_chat.go gating. The /tmp/archie-core-tb3a worktree and its branch are gone, as expected.\n","status":"closed","priority":1,"issue_type":"bug","assignee":"Sam Catlow","owner":"sam@catlow.tech","created_at":"2026-08-13T18:05:35Z","created_by":"Sam Catlow","updated_at":"2026-08-24T07:29:09Z","started_at":"2026-08-13T18:05:39Z","closed_at":"2026-08-13T18:35:26Z","close_reason":"Completed and independently reviewed; worktree left uncommitted per conservative repository policy.","external_ref":"https://github.com/samcharles93/archie-core/issues/485","dependency_count":0,"dependent_count":0,"comment_count":0} {"_type":"issue","id":"archie-core-dzx","title":"Remove duplicate archie-agent safe.directory setup","description":"## Problem\\ncmd/archie-agent/main.go configures git safe.directory inline and then calls markWorktreeSafe, so startup writes the same global git configuration twice.\\n\\n## Scope\\nKeep markWorktreeSafe as the single tested production path and remove the older inline exec.Command block/import. Add regression coverage proving startup wiring has exactly one authoritative safe-directory call.\\n\\n## Acceptance\\n- archie-agent startup invokes markWorktreeSafe exactly once\\n- no second inline safe.directory command remains in main.go\\n- cmd/archie-agent tests and repository quality gate pass","acceptance_criteria":"Single tested safe.directory setup path remains; duplicate inline command is removed; tests and task check pass","notes":"Red proof: TestMainConfiguresWorktreeSafetyOnce failed with 2 setup paths before implementation. Green proof: go test/vet/build ./cmd/archie-agent/... passed; golangci-lint --new-from-rev=HEAD reported 0 issues. Fresh adversarial review found no blockers. Full task check was not run because the working tree contains unrelated in-progress changes and task check rewrites files; scoped package gate used per repository protocol.","status":"closed","priority":1,"issue_type":"bug","assignee":"Sam Catlow","owner":"sam@catlow.tech","created_at":"2026-08-13T16:10:09Z","created_by":"Sam Catlow","updated_at":"2026-08-13T16:13:48Z","started_at":"2026-08-13T16:10:24Z","closed_at":"2026-08-13T16:13:48Z","close_reason":"Removed duplicate inline safe.directory setup; markWorktreeSafe is now the single tested production path","external_ref":"https://github.com/samcharles93/archie-core/issues/470","dependency_count":0,"dependent_count":0,"comment_count":0} +{"_type":"issue","id":"archie-core-v9gv.2","title":"Refactor hyperscript el() usage to Preact JSX","status":"open","priority":2,"issue_type":"task","owner":"sam@catlow.tech","created_at":"2026-09-02T13:37:19Z","created_by":"Sam Catlow","updated_at":"2026-09-02T13:37:19Z","dependencies":[{"issue_id":"archie-core-v9gv.2","depends_on_id":"archie-core-v9gv","type":"parent-child","created_at":"2026-09-02T13:37:19Z","created_by":"Sam Catlow","metadata":"{}"}],"dependency_count":0,"dependent_count":0,"comment_count":0} +{"_type":"issue","id":"archie-core-v9gv.3","title":"Update main entrypoint to mount via Preact render","status":"open","priority":2,"issue_type":"task","owner":"sam@catlow.tech","created_at":"2026-09-02T13:37:19Z","created_by":"Sam Catlow","updated_at":"2026-09-02T13:37:19Z","dependencies":[{"issue_id":"archie-core-v9gv.3","depends_on_id":"archie-core-v9gv","type":"parent-child","created_at":"2026-09-02T13:37:19Z","created_by":"Sam Catlow","metadata":"{}"}],"dependency_count":0,"dependent_count":0,"comment_count":0} +{"_type":"issue","id":"archie-core-v9gv.1","title":"Rename UI view files from .js to .jsx","status":"open","priority":2,"issue_type":"task","owner":"sam@catlow.tech","created_at":"2026-09-02T13:37:18Z","created_by":"Sam Catlow","updated_at":"2026-09-02T13:37:18Z","dependencies":[{"issue_id":"archie-core-v9gv.1","depends_on_id":"archie-core-v9gv","type":"parent-child","created_at":"2026-09-02T13:37:18Z","created_by":"Sam Catlow","metadata":"{}"}],"dependency_count":0,"dependent_count":0,"comment_count":0} +{"_type":"issue","id":"archie-core-v9gv","title":"Migrate Web UI to Preact","status":"open","priority":2,"issue_type":"epic","owner":"sam@catlow.tech","created_at":"2026-09-02T13:37:08Z","created_by":"Sam Catlow","updated_at":"2026-09-02T13:37:08Z","dependency_count":0,"dependent_count":0,"comment_count":0} +{"_type":"issue","id":"archie-core-t2db.8","title":"Dashboard UI for playbook bindings","description":"The t2db.4 backend shipped the CRUD API (POST/GET/PATCH/DELETE /api/bindings + POST /api/bindings/{id}/approve) plus the state machine and HMAC wiring. Operators cannot author bindings from the dashboard yet: a list-and-approve page exists in api_binding.go's response shape but the matching ui/src/ feature folder does not. Build the click-through editor mirroring ui/src/mappings/ (three-file split: bindings.js, binding-editor.js, bindings.css per docs/prds/payload-field-mapping.md:145-152): source matcher, mapping picker, workflow picker, secret input, state-machine controls (create -\u003e edit -\u003e approve -\u003e armed). Also add binding.Owner / binding.Repo fields so multi-repo deployments have a UI to pin a binding to one repo (today the dispatch loop falls back to the single configured repo with a warning; multi-repo is broken). See docs/prds/webhook-intake-security.md and docs/prds/payload-field-mapping.md.","status":"open","priority":2,"issue_type":"feature","owner":"sam@catlow.tech","created_at":"2026-09-02T13:24:13Z","created_by":"Sam Catlow","updated_at":"2026-09-02T13:24:13Z","dependencies":[{"issue_id":"archie-core-t2db.8","depends_on_id":"archie-core-t2db","type":"parent-child","created_at":"2026-09-02T13:24:13Z","created_by":"Sam Catlow","metadata":"{}"}],"dependency_count":0,"dependent_count":0,"comment_count":0} +{"_type":"issue","id":"archie-core-t2db.7","title":"Encrypt binding secrets at rest","description":"binding.Binding.Secret is currently stored as plaintext in the bindings.secret TEXT column. Anyone with read access to the SQLite file (the same access level as captured_events.body and tasks.body) sees HMAC secrets in plaintext. Add at-rest encryption for the secret column so a leaked database file does not leak sender-side shared secrets. See docs/prds/webhook-intake-security.md point 5. Decide the cipher (likely AES-GCM with a key derived from archied's secret registry or a deployment-level key file), the key-rotation story, and whether captured_events.body redaction is upgraded at the same time. Phase F work for t2db.","status":"open","priority":2,"issue_type":"feature","owner":"sam@catlow.tech","created_at":"2026-09-02T13:24:12Z","created_by":"Sam Catlow","updated_at":"2026-09-02T13:24:12Z","dependencies":[{"issue_id":"archie-core-t2db.7","depends_on_id":"archie-core-t2db","type":"parent-child","created_at":"2026-09-02T13:24:12Z","created_by":"Sam Catlow","metadata":"{}"}],"dependency_count":0,"dependent_count":0,"comment_count":0} {"_type":"issue","id":"archie-core-b6ew.6","title":"Inline editing for per-repo config fields (allow_concurrent, max_retries, review_enabled)","description":"archie-core-b6ew.4 shipped visibility for AllowConcurrent, MaxRetries, and ReviewEnabled on the repositories card (read-only), which was the concrete discoverability gap Sam hit. Actual inline editing was scoped out of that issue because it hits a real backend limitation discovered while implementing it: the dashboard's PATCH /api/config path (installUpdateConfigHandler in internal/app/archied/bootstrap.go) applies dotted-key updates via overlay.Nest + configuration.ApplyOverlayValues, and overlay.Nest (internal/infrastructure/configuration/overlay/overlay.go:162) only builds nested maps by splitting on '.' -- it has no array-index handling, so a key like 'repos.0.allow_concurrent' cannot be expressed, and Repos is a []config.Repo slice, not a map, so ApplyOverlayValues's YAML-merge-onto-struct approach cannot target one element either.\n\nThis needs either: (a) a dedicated repo-update seam (e.g. a webui handler keyed by owner/name that reads-modifies-writes the Repos slice directly, bypassing the generic dotted-overlay path), or (b) extending overlay.Nest/ApplyOverlayValues to support array indices generically, which is broader-reaching (touches every list-shaped config field, not just Repos) and needs its own design decision about whether index-based overlay keys are stable across a file edit that reorders [[repos]] entries. Option (a) is almost certainly the smaller, safer change -- decide in a short design note before implementing, per this repo's settled-design-first rule, since option (b) has real footguns (index stability) that deserve a deliberate answer, not a guess.","status":"closed","priority":2,"issue_type":"feature","assignee":"Sam Catlow","owner":"sam@catlow.tech","created_at":"2026-09-02T12:22:53Z","created_by":"Sam Catlow","updated_at":"2026-09-02T12:34:55Z","started_at":"2026-09-02T12:26:59Z","closed_at":"2026-09-02T12:34:55Z","close_reason":"Traced the mechanism before implementing: overlay.Nest/ApplyOverlayValues has no array-index support, but Repos is addressed as a WHOLE dotted key ('repos', no dot) whose value is the full slice -- so no changes to overlay/configuration packages were needed at all. Added PATCH /api/config/repos/{owner}/{name} (webui.Server.UpdateRepoField seam + handleConfigRepoUpdate), wired in archied/bootstrap.go's installUpdateRepoFieldHandler: reads the daemon's own live, untrimmed Repos (not webui.RepoView, which omits Preflight/TestGlob and would silently drop them on a naive round-trip), mutates the one matched repo's one field via the pure applyRepoFieldUpdate helper, and republishes through the existing UpdateConfig(ctx, {\"repos\": repos}) path unchanged -- zero duplicated validate/persist/publish logic.\n\nFrontend: config-row.js's row() gained an opts.save override so a caller can redirect where an edit persists; the repositories table renders an autosaving checkbox (allow_concurrent, review_enabled) and a save-on-blur/Enter text input (max_retries) via api.configRepoUpdate, reverting/erroring inline on failure rather than reloading the whole page mid-edit.\n\nGo tests cover the pure applyRepoFieldUpdate function directly (matched-repo-only mutation, untouched-field preservation including Preflight, type coercion, unknown repo/field rejection) and the webui handler (seam delegation, editable-field allowlist enforced before the seam is called, 503/400 mapping). JS tests cover the structural render (checkbox/input initial state) -- the save round trip itself isn't exercisable through the test shim (addEventListener is a no-op there), consistent with existing config-row.test.js coverage.","dependencies":[{"issue_id":"archie-core-b6ew.6","depends_on_id":"archie-core-b6ew","type":"parent-child","created_at":"2026-09-02T12:22:53Z","created_by":"Sam Catlow","metadata":"{}"}],"dependency_count":0,"dependent_count":0,"comment_count":0} {"_type":"issue","id":"archie-core-c0y4","title":"TelegramConfig gains a bws-capable Token (SecretRef), not just TokenEnv","description":"internal/config/config.go's TelegramConfig only has TokenEnv (a plain OS env var name); it never resolves through secret.Registry the way Forge.Token and Provider.APIKey do (secret.SecretRef{engine,key}), per ARCHITECTURE.md:220-224's own note that TelegramConfig.TokenEnv is one of the older env-only fields kept for backwards compatibility. internal/app/archied/telegram_setup.go:75 does a bare os.Getenv(cfg.Chat.Telegram.TokenEnv), bypassing the secret registry entirely.\n\nThis means a bws-stored Telegram token cannot be referenced directly in config.toml the way [providers.openai].api_key or [forge.token] already can -- the operator has to bridge it into a real process env var by hand (bws secret get ... | export ARCHIE_TELEGRAM_TOKEN=...) instead of writing [chat.telegram.token] = { engine = \"bws\", key = \"...\" }.\n\nFix: add a Token secret.SecretRef field to TelegramConfig, mirroring Forge.Token's shape and resolution path. Keep TokenEnv working as a fallback/legacy path (same pattern documented for Provider.APIKeyEnv vs APIKey) rather than a breaking config change -- resolve Token first if set, fall back to TokenEnv. Update config.example.toml and docs/architecture identity/config docs to show the SecretRef form as the preferred one.","status":"open","priority":2,"issue_type":"feature","owner":"sam@catlow.tech","created_at":"2026-09-02T12:21:10Z","created_by":"Sam Catlow","updated_at":"2026-09-02T12:21:10Z","dependency_count":0,"dependent_count":0,"comment_count":0} {"_type":"issue","id":"archie-core-b6ew.4","title":"Structured configuration editors: repositories, maps","description":"Repositories currently render as a static read-only table missing AllowConcurrent, MaxRetries, and ReviewEnabled entirely (ReviewEnabled is absent from RepoView in api_config.go, not just unrendered). Add ReviewEnabled to RepoView, and build a dedicated backend-defined editor for the repositories section (and the models/providers maps) so these stop being silently read-only. This is the concrete fix for the allow_concurrent discoverability gap Sam hit, delivered as a typed editor rather than a one-off card patch, since the schema this epic introduces is what should own its shape.","status":"closed","priority":2,"issue_type":"feature","assignee":"Sam Catlow","owner":"sam@catlow.tech","created_at":"2026-09-02T12:00:46Z","created_by":"Sam Catlow","updated_at":"2026-09-02T12:23:52Z","started_at":"2026-09-02T12:21:13Z","closed_at":"2026-09-02T12:23:52Z","close_reason":"Scoped to visibility, not full inline editing -- while implementing, found that the dashboard's PATCH /api/config path (overlay.Nest + configuration.ApplyOverlayValues) has no array-index support, so a per-element update to the Repos slice cannot be expressed as a dotted key today. That's real, separate scope, filed as archie-core-b6ew.6 rather than bolted on.\n\nShipped: RepoView.ReviewEnabled (previously absent from the API entirely, not just unrendered) alongside AllowConcurrent and MaxRetries; the repositories card now shows Concurrent/Max retries/Self-review columns as read-only pills, closing the concrete discoverability gap (Sam couldn't see allow_concurrent without an agent tracing Go source). config.example.toml documents review_enabled, which was undocumented before this. Go test pins ReviewEnabled reaching ConfigView for both a configured and a zero-value repo.","dependencies":[{"issue_id":"archie-core-b6ew.4","depends_on_id":"archie-core-b6ew","type":"parent-child","created_at":"2026-09-02T12:00:46Z","created_by":"Sam Catlow","metadata":"{}"},{"issue_id":"archie-core-b6ew.4","depends_on_id":"archie-core-b6ew.2","type":"blocks","created_at":"2026-09-02T12:00:51Z","created_by":"Sam Catlow","metadata":"{}"}],"dependency_count":1,"dependent_count":1,"comment_count":0} diff --git a/.beads/metadata.json b/.beads/metadata.json index 0418c21db..42a339e23 100644 --- a/.beads/metadata.json +++ b/.beads/metadata.json @@ -3,5 +3,6 @@ "backend": "dolt", "dolt_mode": "embedded", "dolt_database": "archie_core", - "project_id": "e8a66150-1f5f-42d2-8653-dcf8463cef74" + "project_id": "e8a66150-1f5f-42d2-8653-dcf8463cef74", + "last_bd_version": "1.2.2" } \ No newline at end of file