Applies the grounding discipline from the agentic briefing work to the
main chat system prompt. The regular chat pipeline was already agentic
(it uses stream_chat_with_tools), but its system prompt never told the
model "only assert facts from tool results" or "if a tool returns
nothing, say so honestly." That left room for the same class of
hallucinations the briefings had — calling list_events, getting an
empty array, and then confidently mentioning a meeting anyway.
Adds two new static rules to the tool guidance block in llm.build_context:
GROUNDING — when the user asks about their own data, call the relevant
tool to see what exists. Never assert from memory or assumption.
HONESTY WHEN EMPTY — if a tool returns empty results, tell the user
plainly. No fabricated example items, no invented meetings, no generic
suggestions dressed up as real data.
Both rules are in the static (KV-cache-stable) portion of the system
prompt so they cost nothing on repeated requests for the same user.
Carries the hallucination fix from the briefing work directly into
every chat turn, not just chat that happens inside a briefing thread.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
PR 1.5 (commit 4168167) changed run_slot_injection's signature from
returning str to returning tuple[str, dict] so the scheduler can get
at the agentic message list for receipt persistence. The
test_run_slot_morning_runs_on_work_day mock was still returning the
old plain-string value, which unpacked to a ValueError at runtime.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Manual bump because the pre-commit hook that normally handles this
didn't run — its scripts had lost their executable bits earlier in
the session. Permissions have been restored in this commit's tree,
so subsequent commits touching fable-mcp will auto-bump again.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Adds the MCP tools needed to debug the agentic briefing pipeline
without waiting for scheduled slots to fire:
- fable_list_briefings — list briefing conversations newest first
- fable_get_today_briefing — fetch today's briefing with all messages
- fable_get_briefing_messages(conv_id) — message list for a specific
briefing conversation, including tool_calls with embedded results
- fable_trigger_briefing(slot) — manually run a slot via
POST /api/briefing/trigger (fires RSS/weather refresh, same as
the scheduler)
- fable_get_conversation(conv_id) — generic conversation read for
chat or briefing threads, full messages + tool_calls + metadata
All of these hit existing REST endpoints (/api/briefing/conversations,
/api/briefing/conversations/<id>/messages, /api/briefing/trigger,
/api/chat/conversations/<id>) so no API surface changed — the gap
was purely on the MCP side. Bearer token auth works across both
blueprints because login_required already accepts API keys.
Enables a full "trigger → inspect → tune prompt → repeat" loop from
Claude without touching the UI, which is necessary for iterating on
agentic briefing prompts when the scheduled slot is hours away.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
PR 1.5 + PR 2 of the agentic briefing rollout. Extends the agentic
path to cover the morning/midday/afternoon check-in slots in addition
to the 4am compilation, and persists the full tool-call sequence
into the briefing conversation so chat follow-ups see the receipts.
run_slot_injection now honors the briefing_mode setting the same way
run_compilation does: agentic mode routes through run_agentic_briefing
with a check-in system prompt variant, legacy falls back automatically
if the new path returns empty. Signature changed from str to
(str, dict) to surface the agentic message list to the caller.
_agentic_system_prompt now takes the user's local-day window and
timezone, pre-computed by run_agentic_briefing, and embeds them in
the prompt. This eliminates a whole class of "wrong day" bugs that
would otherwise happen when the model tried to translate "today" into
ISO 8601 ranges without knowing the user's timezone.
briefing_scheduler._run_slot_for_user now calls _persist_agentic_messages
after each slot, which walks the agentic message list and stores
every intermediate assistant turn (with its tool calls and results
folded into the flat storage format the existing chat loader expects)
as a real message row. The synthetic "[Midday briefing update]"
user-role messages are no longer created — final prose is written
with metadata.briefing_slot so the UI can still identify it as a
scheduled entry. The agentic user-trigger ("Generate my morning
briefing…") is deliberately skipped so it doesn't recreate the same
fake-user-message problem we're trying to remove.
briefing_conversations.post_message now accepts tool_calls, matching
the schema the Message model already supports. This lets scheduled
briefings write structured tool-use history without reaching into
the model layer.
Net effect with briefing_mode="agentic" on:
- All four slots are grounded in tool results, no more hallucinated
events or tasks
- Chat follow-ups in the briefing conversation see morning's
list_events → [] receipt (and everything else), so "what meeting?"
gets an honest "nothing on the calendar" reply grounded in data
- No more fake [Briefing update] user messages in the chat scroll
Still to come (PR 3): UI polish — tool-call status row, inline cards
for list_events/list_tasks results, visual treatment for slot messages.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
services/events.py::list_events treated any event with end_dt IS NULL
as perpetually matching, so a point-event created last week would be
returned as "happening today" whenever the briefing or list_events
tool queried today's window. That manifested as briefings claiming
past events were scheduled for the current day.
Split the null-end branch: point events now only match when start_dt
falls within [date_from, date_to]. Events with an end_dt continue to
use standard overlap logic.
This bug affected both the legacy briefing synthesis and the agentic
list_events tool path, since both call the same events service.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
First cut of the agentic briefing redesign. Morning compilation can now
route through a tool-call loop that grounds every factual claim in an
actual tool result, eliminating the hallucinated meetings, tasks, and
news items the legacy one-shot path was producing. Behind a per-user
`briefing_mode` setting (default "legacy"); falls back to the legacy
path automatically if the new path returns empty (e.g. model too weak
to drive tool calls reliably).
New: services/briefing_tools.py — explicit read-only allowlist of 10
tools (tasks, events, weather, rss, projects, notes). New tools added
to tools.py must be opted in by name. Excludes all mutating tools and
external search tools (search_images, search_web, research_topic) which
are neither useful nor safe for a scheduled background job.
New: briefing_pipeline.run_agentic_briefing — wraps the existing
stream_chat_with_tools loop with slot-specific system prompts that tell
the model to only assert facts from tool results and to be honest when
tools return nothing. Max 8 rounds, per-round exception handling,
returns the full message list so tool-call receipts can be persisted
alongside the prose in a later PR.
Sentence-count floors bumped: compilation 6–10 (was 4–8), check-ins
3–5 (was 2–3). Weekly review 5–8.
Design doc: docs/2026-04-10-agentic-briefing-design.md
Out of scope for this PR (future PRs): slot-injection migration,
persisting tool-call receipts into the conversation so chat follow-ups
see them, UI polish for tool-call status, sidecar storage for
briefings. See the design doc's migration path for details.
Enable on an account with:
UPDATE settings SET value='agentic'
WHERE user_id=<id> AND key='briefing_mode';
or insert the row if missing.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace the hardcoded "2h" keep_alive everywhere with a helper that
returns OLLAMA_KEEP_ALIVE_MAIN (default 30m) for the interactive model
and OLLAMA_KEEP_ALIVE_BACKGROUND (default 10m) for the background
model. Lets the main model release VRAM during long idle periods
while keeping it warm enough for bursty chat use, and stops the
sporadic background model from camping VRAM it rarely needs.
Seven call sites updated to route through llm.keep_alive_for(model):
the streaming helpers, generate_completion, the two startup warmers,
the settings KV-cache primer, and the chat warmer endpoint.
Override via env vars: OLLAMA_KEEP_ALIVE_MAIN, OLLAMA_KEEP_ALIVE_BACKGROUND.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Replace the hand-edit-JSON instructions in Settings → API Keys with a
tabbed UI (Claude Code / Claude Desktop / Other). The Claude Code tab
leads with the `claude mcp add` command, pre-filled with FABLE_URL and
the most recently generated API key, plus copy-to-clipboard buttons on
every snippet. Recommend `uv tool install` or `pipx install` over bare
`pip install` so fable-mcp reliably lands on PATH under PEP 668.
Also fix incorrect priority values in fable-mcp tool docstrings — the
enum is `none|low|medium|high`, not `low|normal|high`.
DRY pass: extract shared `copyToClipboard()` helper used by both
`copyApiKey` and the new snippet buttons; reuse existing
`btn btn-secondary btn-sm` for the copy buttons instead of a bespoke
class.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Even mode=min still hits the registry's blob size limit (400 Bad Request).
The local runner's Docker daemon layer cache is sufficient for fast
incremental builds without needing a separate registry cache tag.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
mode=max exports all intermediate layer blobs which exceeded the Forgejo
registry's blob upload size limit (400 Bad Request). mode=min only exports
the final image layers, keeping cache entries small enough for the registry.
The actual image build and push was succeeding; only the cache write failed.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
.logo-book fill was var(--color-text) which renders black in light mode.
Changed to var(--color-primary) so it matches the brand indigo in both
modes. Stroke updated to match at 70% opacity.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Design direction: "Illuminated Transcript" — the AI response is the
primary artifact; user prompts are subordinate margin notes.
ChatMessage.vue:
- User bubble: ghosted background (rgba 3-4% opacity), muted text,
thin near-invisible border. De-emphasized, recessed.
- Assistant bubble: left indigo accent border (2px), elevated shadow
via --color-bubble-asst-shadow. Commands attention.
- Assistant role label: indigo (var(--color-primary)); user label muted.
- Removed old white-override content rules (no longer on dark bg).
- Thinking block: full border → left-only indigo accent bar (35% opacity),
removed inner top border.
ChatView.vue:
- Streaming bubble: same left accent + shadow as assistant bubbles.
- Streaming role label: indigo to match ChatMessage.
- Thinking block in streaming: same treatment as ChatMessage.
theme.css:
- Added --color-bubble-user-bg/border/text for light + dark.
- Added --color-bubble-asst-shadow for light + dark.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Replaces docker-ce-cli (requiring external Docker apt repo) with
docker.io from Ubuntu 24.04 main, which is sufficient for docker login
and working alongside buildx.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Single combined layer was ~280MB and consistently failed to push with
499 (nginx closing connection). Split into three layers (~70-100MB each)
so each is pushed independently and stays within the timeout.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
docker/login-action requires the docker CLI binary in the job container.
The daemon is provided by the host socket mount; only the CLI is needed.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Ubuntu 24.04 blocks pip install into the system Python by default.
Add --break-system-packages for the lint job's throwaway container.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Labels are defined in act-runner-config.yml which is the authoritative
source. The env var is only used on first registration and was being
ignored after .runner was created.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
The label now explicitly declares the dependency versions it provides
rather than using the opaque ubuntu-latest alias. All runs-on fields
in ci.yml updated to match.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Move ubuntu-latest → runner-base:py3.12-node22 mapping into
act-runner-config.yml so it is respected at runtime. The
GITEA_RUNNER_LABELS env var is only honoured on first registration
and was being ignored after the runner's .runner file was created.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
- infra/Dockerfile.runner-base: new Ubuntu 24.04 image with Python 3.12
and Node 22 LTS pre-installed, tagged py3.12-node22. Eliminates the
~3-4 min deadsnakes PPA install that ran on every test job.
- infra/runner-compose.yml: update ubuntu-latest label to use
runner-base:py3.12-node22 instead of node:20-bullseye.
- .forgejo/workflows/ci.yml: remove Python 3.12 install steps (now in
base image), add actions/cache for pip keyed on pyproject.toml hash,
bump node-version 20 → 22.
- Dockerfile: bump frontend build stage node:20-alpine → node:22-alpine.
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Three-layer guard on create_note and create_task:
1. Exact title match (get_note_by_title, case-insensitive) → hard block,
redirects model to use update_note instead
2. Fuzzy title match (SequenceMatcher ≥82%, punct-stripped word search for
candidates) → hard block; catches "Game Premise" vs "Game Premise Notes"
3. Semantic content similarity (semantic_search_notes threshold=0.87) → soft
block with requires_confirmation:true; model asks user then retries with
confirmed:true; graceful no-op if embedding model is unavailable
create_project and create_milestone now always require confirmed:true before
creation (structural entities — must be intentional). Both are also guarded
by exact + fuzzy title checks post-confirmation.
create_note and create_task gain an optional confirmed parameter (not
required by default; only used to bypass a content-similarity soft-block
after the user has been asked).
Helpers added to tools.py:
- _fuzzy_title_match(title, candidates, threshold=0.82)
- _PUNCT_RE for stripping punctuation before word-search candidate fetch
- difflib.SequenceMatcher and re imported at module level
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
Backend:
- routes/tasks.py: POST + PUT were silently dropping project_id,
milestone_id, parent_id from request body — root cause of association
not saving from the task editor
- routes/tasks.py: GET /api/tasks/:id now includes parent_title in
response (secondary lookup when parent_id is set)
- routes/notes.py: add PATCH /api/notes/:id for partial updates (used
by sub-task status toggle; PUT already existed but PATCH was missing)
- routes/projects.py: GET /api/projects/:id/notes now fetches milestone
IDs and passes them via milestone_ids so tasks assigned to a milestone
(but lacking project_id) are included in the project view
- services/notes.py: create_note() auto-sets project_id from milestone
when milestone_id is provided and project_id is omitted; list_notes()
gains milestone_ids param — when combined with project_id uses OR
condition (project_id=X OR milestone_id IN (...))
Frontend:
- NoteEditorView: add MilestoneSelector; milestone resets when project
changes; all save paths (save/create/auto-save) include milestone_id
- stores/notes.ts: add milestone_id to createNote + updateNote types
- TaskEditorView: sub-tasks now inherit milestone_id from parent task
- AppHeader: three-zone layout — brand left, Notes/Projects/Tasks/Chat
centered (absolute positioning), right rail with status/theme/? and
gear dropdown containing Settings/Users/Logs; mobile dropdown unchanged
Co-Authored-By: Claude Sonnet 4.6 <noreply@anthropic.com>
