db-context-supabase

Output contract, in brief. This body is over the per-skill re-injection cap, so after a compaction the sections below are truncated away while this summary survives. They remain authoritative in full; re-read this file before emitting if you need them.

• Standard output layout (required): Produce the command output using this structure (English only):
• Artifact changes: Follow ## Global output contract in WORKFLOW_OPERATING_SYSTEM.md for APPLIED / PROPOSED / SKIP rules. Every listed file...
• Command transcript: Brief audit trail (max 4 lines; max 3 in no-op runs with NO_OP_TRACE).
• Handoff: Use the adaptive ending format of WORKFLOW_OPERATING_SYSTEM.md ## Global output contract. Every Handoff is one fenced text b...
• Definition of done (command output): The MCP precondition check is performed first and its result is reported. When it fails, the local CLI path is checked next; when...

Act as a senior/staff engineering database context capture for the active task, scoped to Supabase.

Goal: Validate that a Supabase MCP server is configured and reachable, introspect a user-scoped subset of the database (tables, columns, types, RLS policies, and optionally functions/migrations), and persist the result as DB_CONTEXT.md inside the active task folder so the task has a grounded, point-in-time schema reference for planning, implementation, and review.

When no MCP server is reachable, the command falls back to the local Supabase CLI and produces the same snapshot from CLI output. Only when neither path is available does it end in a no-op.

This command is opt-in. It is not part of the default task initialization flow; run it after task-init only when the task actually touches Supabase data, schema, or RLS.

Mandatory context bootstrap (before any output):

• Resolving the spec and wos/<topic>.md: canonical repository FIRST. Otherwise use active-adapter docs: ~/.claude/workflow-docs (CLAUDE_WORKFLOW_DOCS_DIR), ~/.cursor/workflow-docs (WORKFLOW_DOCS_DIR; project .cursor/workflow-docs on project installs), ~/.codex/workflow-docs (CODEX_WORKFLOW_DOCS_DIR), or <kimi-home>/workflow-docs (KIMI_WORKFLOW_DOCS_DIR). Nonempty overrides win. Kimi home: KIMI_CODE_HOME, else existing ~/.kimi-code, else existing ~/.kimi, default ~/.kimi-code. Use one root for both paths; try other adapters if absent or unknown. Report resolved root, fallback to other adapters, or missing mandatory context in ### Command transcript. Installed snapshots refresh and prune retired topics only on sync.
• Read these sections in WORKFLOW_OPERATING_SYSTEM.md first:
  • ## LLM execution contract
  • ## Editor mode policy
  • ## Global output contract (including Adaptive handoff and Mode selection rule)
  • ## Cross-cutting workflow guardrails
  • ## Evidence priority
• Read additional sections only when needed:
  • naming/path setup: ## Naming conventions, ## Repository structure
  • artifact requirements: ## Task files, ## TASK_STATE policy
• Read the active task's TASK_STATE.md to confirm there is an active task to attach the DB context to and to align the requested table/schema scope with the task objective.
• Read the active task's SOURCE_OF_TRUTH.md to detect any prior ## DB context link before proposing a new one.
• Read canonical command inventory when a clone is available; otherwise use installed Fhorja commands and skills.

Required inputs:

• active task path (or enough context to resolve projects/<client>__<project>/active/YYYY-MM-DD_<task-slug>/)
• scope of introspection, supplied by the user as one or more of:
  • list of table names (qualified schema.table or unqualified for the default public schema)
  • list of schemas to include in full (use sparingly; large schemas should be narrowed to specific tables)
• depth flag (one of):
  • tables-only: tables and columns with types, primary keys, and not-null flags
  • tables+rls (default): tables-only plus row-level security policies for each in-scope table
  • full: tables+rls plus relevant functions, triggers, and the most recent migration filenames touching in-scope tables
• optional: Supabase project ref (if the MCP server is configured for multiple projects); when omitted, use the project the MCP server is currently bound to and record which one was used
• optional: refresh flag (refresh to regenerate an existing DB_CONTEXT.md; default is to fail with NO_OP_TRACE if a non-stale DB_CONTEXT.md already exists for the same scope)
• optional: path to the local Supabase project directory. Needed only on the local CLI path below, where every CLI form runs from that directory. When omitted and the MCP path is unavailable, ask for it once instead of guessing a directory.

Project repository files to read:

• projects/<client>__<project>/active/YYYY-MM-DD_<task-slug>/TASK_STATE.md
• projects/<client>__<project>/active/YYYY-MM-DD_<task-slug>/SOURCE_OF_TRUTH.md
• projects/<client>__<project>/active/YYYY-MM-DD_<task-slug>/DB_CONTEXT.md (only if it already exists, for refresh comparison)

Project repository files to update:

• projects/<client>__<project>/active/YYYY-MM-DD_<task-slug>/DB_CONTEXT.md (create or fully regenerate; never partial-merge)
• projects/<client>__<project>/active/YYYY-MM-DD_<task-slug>/SOURCE_OF_TRUTH.md (append-only: add a single ## DB context section pointing to ./DB_CONTEXT.md if not already present)

Introspection path selection (MCP first, then the local CLI): Choose the path before any introspection, in this order, and record which one was used.

1. Supabase MCP server. When the MCP server is configured, reachable, and exposes introspection tools, use it. Nothing about this path changes: same read-only rules, same scope and depth handling, same output.
2. Local Supabase CLI. WHEN no MCP server is reachable AND a local Supabase CLI is available, fall back to the CLI and produce DB_CONTEXT.md from its output under the local CLI branch rules below.
3. Neither. WHEN no MCP server is reachable and no local CLI is available, propose no DB_CONTEXT.md content and end with NO_OP_TRACE plus one actionable configuration line naming both paths (configure the Supabase MCP server, or bring the local stack up with supabase start in the local project directory and re-run). This is the existing no-op behavior and it is unchanged.

Local CLI branch (applies only when path 2 was selected):

• Availability check first. It reads the state of the local stack, not the database. Run supabase status from the local Supabase project directory. When it exits non-zero or reports the stack as not running, propose no DB_CONTEXT.md content and end with NO_OP_TRACE plus the actionable line naming supabase start. This command never starts, resets, seeds, or pushes to the stack; the user runs supabase start and re-runs this command.
• Secrets rule, non-negotiable. supabase status prints the JWT secret, the anon key, and the service_role key to stdout. That raw output is NEVER written into DB_CONTEXT.md, NEVER into SOURCE_OF_TRUTH.md, NEVER into the command transcript or any other task artifact, and NEVER echoed back in the chat output. Record three facts and nothing more: that the local stack is reachable, the DB host, and the DB port, the last two read off the DB URL line with the user and password components stripped (the captured example postgresql://postgres:postgres@127.0.0.1:54322/postgres becomes host 127.0.0.1 and port 54322). Pasting a status block into an artifact leaks a service-role credential, which bypasses RLS, into task memory that outlives the run.
• Read-only, restated in full here because it must not live only on the MCP branch. On this branch, never run INSERT / UPDATE / DELETE / DROP / ALTER / TRUNCATE / GRANT / REVOKE through the CLI or through any connection the CLI reports. If the user requests one, redirect them to the appropriate implementation command and stop.
• The read-only set for this branch is exactly these forms, each cited to the captured Supabase CLI reference, and nothing outside it may be issued:
  • supabase status (availability only, subject to the secrets rule above)
  • supabase db dump --local -f <path> for the schema
  • supabase db diff --local
  • supabase db lint --local
  • supabase inspect --local db <bloat|blocking|calls|locks|outliers>
  • supabase inspect --local report [--output-dir <string>] Any subcommand outside this list is out of scope for this command, including anything that starts, resets, seeds, pushes, or applies migrations.
• Write the schema dump to a scratch path outside the task folder, read it there, and do not add it as a task artifact. DB_CONTEXT.md plus the single ## DB context cross-link stay the only files this command touches.
• Target selection is explicit. The official CLI command reference (checked 2026-09-07) documents --local for dump, diff, lint, and inspect; dump and inspect otherwise default to the linked project. Run every listed operation from the same project directory checked by supabase status, with --local as shown. --linked and --db-url are not allowed on this local branch. Help for these listed subcommands may be read to verify installed-version support; if the local selector is unsupported or the target cannot be established, stop with NO_OP_TRACE and name the missing CLI or target evidence instead of emitting a snapshot labeled local. Narrow to in-scope tables while reading the dump. Record the CLI version only from observed output, otherwise keep [unclear in source] for the version.
• Any field the CLI output does not answer (an RLS policy, a function body, a migration filename) is marked [unclear from CLI output], never guessed.

Operating rules:

• Do not implement production code, migrations, or destructive SQL.
• Only issue read-only introspection: through the MCP server on the MCP path, and through the read-only CLI forms listed above on the local CLI path. Never run INSERT / UPDATE / DELETE / DROP / ALTER / TRUNCATE / GRANT / REVOKE for any reason on either path; if the user requests one, redirect them to the appropriate implementation command and stop.
• Never record a credential. No supabase status output block, no service_role key, no anon key, no JWT secret, and no full connection string appears in DB_CONTEXT.md, in SOURCE_OF_TRUTH.md, in the transcript, or in the chat output. Host and port are the only connection details worth recording.
• Do not invent tables, columns, types, policies, triggers, functions, or migrations. Every recorded field must come from the MCP server's response, from the CLI output on the local path, or from explicit user-supplied scope.
• Always record Last refreshed: as today's date in YYYY-MM-DD format, the introspection path used, and the resolved Supabase project identifier (local on the local CLI path). Stale snapshots without these fields are invalid.
• Never dump the entire database. If the user provides a schema with more than 25 tables and no narrowing list, ask one targeted question to narrow scope before introspecting.
• Re-run policy: regeneration replaces DB_CONTEXT.md in full. Do not partial-merge. If the user has handwritten notes that should survive refresh, those belong in DECISIONS.md or TASK_STATE.md, not in DB_CONTEXT.md. State this explicitly when proposing a refresh that overwrites an existing file.
• Cross-link policy: SOURCE_OF_TRUTH.md gets at most one ## DB context section with a single relative pointer to ./DB_CONTEXT.md. Do not duplicate schema content into SOURCE_OF_TRUTH.md.
• Do not modify other task-scoped artifacts (TASK_STATE.md, DECISIONS.md, IMPLEMENTATION_PLAN.md, IMPACT_ANALYSIS.md, INVARIANTS_AND_NON_GOALS.md, TEST_STRATEGY.md, PR_PACKAGE.md, slice files).
• Treat task-memory write policy per WORKFLOW_OPERATING_SYSTEM.md: write the file and mark it APPLIED.
• Mark fields as [unclear from MCP response] on the MCP path, or [unclear from CLI output] on the local CLI path, when the introspection returned ambiguous or partial data; never fill gaps with guesses.
• Output is intentionally bounded. Do not produce schema analysis, ER diagrams, or design recommendations beyond the snapshot itself; routing those follow-ups belongs in subsequent commands (impact-analysis, decision-interview, implementation-plan).
• Official next-command names only: every recommended next command (including the handoff Run now line) MUST use a registered canonical command name or available installed Fhorja command or skill. Never invent names. One exception: Run now: none with Mode: N/A declares that the chain has ended and no following command would be honest, defined under ### Official command names (routing integrity) (ADR-0126); use it only when nothing honest remains, never to end a chain that has a real next step.
• Handoff: end with the adaptive ### Handoff block per WORKFLOW_OPERATING_SYSTEM.md ## Global output contract (Mode A compact or Mode B full). Default Run now: read TASK_STATE.md Last completed step; if it was task-init, default to impact-analysis; if uncertain, default to what-next.

Snapshot format (canonical):

# DB_CONTEXT

## Snapshot metadata
- Provider: supabase
- Introspection path: mcp | local-cli
- Project ref: <resolved-project-ref-or-alias, or "local" on the local-cli path>
- Local stack: reachable (local-cli path only; never a key, a secret, or a connection string)
- DB host/port: <host>:<port> (local-cli path only; host and port only, no user, no password)
- CLI version: <version if the CLI output reported one, else [unclear in source]> (local-cli path only)
- Last refreshed: YYYY-MM-DD
- Depth: tables-only | tables+rls | full
- Scope: <comma-separated list of schema.table or schema.* entries actually introspected>

## Tables

### <schema>.<table>
- Columns:
  - <column_name> <type> [PK] [NOT NULL] [DEFAULT <expr>]
- Primary key: <column(s)>
- Foreign keys:
  - <column> -> <ref_schema>.<ref_table>(<ref_column>)
- Indexes (non-PK, non-unique-constraint):
  - <index_name> ON (<columns>)
- RLS enabled: yes | no
- RLS policies (only when depth >= tables+rls):
  - <policy_name> [<command: SELECT/INSERT/UPDATE/DELETE/ALL>] FOR <role>: <USING expression> / <WITH CHECK expression>

## Triggers (only when depth = full)
- <trigger_name> on <schema>.<table>: <trigger definition from the MCP response or CLI output>

## Functions (only when depth = full)
### <schema>.<function_name>(<args>)
- Returns: <return_type>
- Language: <plpgsql | sql | ...>
- Security: <DEFINER | INVOKER>
- One-line summary: <verbatim COMMENT ON FUNCTION if present, else "[no COMMENT in source]">

## Recent migrations touching in-scope tables (only when depth = full)
- <migration_filename_or_id>: <one-line description from migration metadata, or "[no description in source]">

Sections that have no content for the chosen depth must be omitted entirely rather than left empty.

Required output:

1. Resolved active task path.
2. Which introspection path was selected (mcp, local-cli, or none) and the result of its precondition check. For the MCP path: configured / reachable / introspection tools available, or the actionable failure line. For the local CLI path: whether the local stack is reachable, reported without any key, secret, or connection string.
3. Resolved Supabase project ref/alias used for introspection, or local plus the DB host and port on the local CLI path.
4. Resolved scope (tables/schemas) and depth flag.
5. Whether this is a create or a refresh of DB_CONTEXT.md, and (on refresh) a one-line drift summary versus the prior snapshot (e.g., "3 new columns, 1 dropped table, RLS toggled on public.orders").
6. Exact content for DB_CONTEXT.md using the canonical snapshot format.
7. Exact patch to SOURCE_OF_TRUTH.md adding the ## DB context cross-link, or SKIP if the link is already present.
8. Recommended next command (must exist as commands/<name>.md or commands/<name>/SKILL.md; verify against directory listing before output). In installed mode use the available Fhorja inventory.
9. Recommended editor mode for that next command.
10. Why that is the correct next step.

Claim grounding (active epistemic humility)

Claim grounding (active epistemic humility). This block governs what you may assert and how you record it. It is keyed to the substrate section you are writing, not to which command is running, and it is INERT on any output that writes none of the claim-bearing sections below. Full contract and rationale: wos/active-epistemic-humility.md.

1. When this applies. This block fires ONLY while you are writing a claim-bearing substrate section: TASK_STATE.md ## Current known facts, ## Risks to watch, ## Observations, ## Active files in scope, ## Canonical decisions; DECISIONS.md ## Locked decisions; IMPLEMENTATION_PLAN.md ## Current gaps, ## Risks and mitigations; IMPACT_ANALYSIS.md; EXTERNAL_RESEARCH.md; REFERENCES.md; or any section whose content is a statement a later command or a human decision will act on. WHEN your output writes none of these, this block imposes nothing: skip it and proceed. This is the D-13 inert clause; a fully-grounded or claim-free output pays nothing.
2. The unit is the load-bearing claim. A load-bearing claim is one a downstream command or a human decision consumes. A passing aside is not load-bearing; a statement someone will act on is. Apply the rest of this block per load-bearing claim, not per sentence.
3. Ground it or abstain. Before you assert a load-bearing claim, trace it to the enumerable grounded set: a captured REFERENCES.md entry, a file read in this session, command output actually seen, or a passing deterministic gate. A claim supported only by model memory is OUTSIDE the grounded set, including when you are right, because that support is not observable. WHEN a load-bearing claim falls outside the set, do NOT assert it: either investigate until it is grounded, or abstain per rule 6.
4. Status records provenance, never confidence. WHERE you attach an epistemic status to a claim, the status names WHERE THE CLAIM CAME FROM: a REFERENCES.md entry title, a file path plus line, or the gate output it came from. It SHALL NOT express a degree of certainty. Do NOT add a confidence field, a numeric threshold, or a self-assessment prompt anywhere; a self-reported confidence signal is not a usable control signal (wos/active-epistemic-humility.md Part 1.3). A status whose referent slot is empty is read as UNKNOWN, not as a weak yes.
5. Persisted claims carry the status; chat-only claims carry it when they route. Every load-bearing claim you write into a task-memory artifact carries its provenance referent, and that referent travels with the claim so a later command reads it too; do not drop it at the write boundary. A load-bearing claim that appears only in a chat-turn output carries a status only when it crosses the grounding boundary and triggers a route (an abstention, an escalation).
6. Abstain as a routed continuation, never a bare refusal. WHEN you abstain, name the specific investigation that would settle the question AND route to the command that runs it (capture-references, code-locate, incident-triage, or the fitting one). A withholding that stalls the work is invalid output. Abstention is distinct from NO_OP: NO_OP means there is no work to do; abstention means there is work and the grounding to do it is missing.
7. An unfired gate is not evidence. The absence of a fired check does not mean grounding existed. Do not read silence here as a pass.

Standard output layout (required)

Produce the command output using this structure (English only):

Artifact changes

Follow ## Global output contract in WORKFLOW_OPERATING_SYSTEM.md for APPLIED / PROPOSED / SKIP rules. Every listed file carries one of those three tokens, in Lean output too; a prose verb like "written" is not a label.

Command transcript

Brief audit trail (max 4 lines; max 3 in no-op runs with NO_OP_TRACE).

Handoff

Use the adaptive ending format of WORKFLOW_OPERATING_SYSTEM.md ## Global output contract. Every Handoff is one fenced text block with all four lines, Run now:, Mode:, Work complexity: and Reason:, on a stop and on a refusal too; the terminal form is Run now: none with Mode: N/A. A new Mode:, a model or a fresh session is never a stop, an offered choice is one, and Reason: names a role, never a model.

Definition of done (command output)

• The MCP precondition check is performed first and its result is reported. When it fails, the local CLI path is checked next; when neither path is available, no DB_CONTEXT.md content is proposed and the run ends with NO_OP_TRACE plus the actionable configuration line.
• The proposed DB_CONTEXT.md includes Provider, Introspection path, Project ref, Last refreshed, Depth, and Scope metadata, and at least one populated ## Tables entry (or an explicit NO_OP_TRACE if the requested scope returned no tables).
• Every column type, policy, trigger, function, and migration entry is grounded in the MCP server's response or in the CLI output; no field is fabricated. Ambiguous fields are marked [unclear from MCP response] or [unclear from CLI output].
• No destructive SQL was issued on either path. Only read-only introspection was used, and on the local CLI path only the command forms in that branch's read-only set.
• No credential escaped the local CLI path: no supabase status output block, no service_role key, no anon key, no JWT secret, and no full connection string appears in DB_CONTEXT.md, in SOURCE_OF_TRUTH.md, in the transcript, or in the chat output.
• No task-scoped artifact other than DB_CONTEXT.md and (at most) a single ## DB context cross-link in SOURCE_OF_TRUTH.md is modified.
• On refresh, the prior DB_CONTEXT.md is fully replaced (no partial merge), and the drift summary in the required output makes the change auditable.
• ### Artifact changes marks the patches as APPLIED.
• The basename in the Run now: line corresponds to a real file in commands/<name>.md or commands/<name>/SKILL.md. In installed mode use the available Fhorja inventory.
• Output ends with a complete ### Handoff block per the adaptive format in WORKFLOW_OPERATING_SYSTEM.md ## Global output contract.
• Before declaring this output done, confirm it satisfies the shared Definition of done (command outputs) and Gate conditions in WORKFLOW_OPERATING_SYSTEM.md.

Quality bar: Optimize for fidelity to the live Supabase schema, point-in-time auditability, narrow scope (no full-DB dumps), credential hygiene (a service-role key never reaches a task artifact), and minimal disruption to whatever task-scoped work was in progress before this capture.