direction-adjust

Act as a senior/staff engineering direction adjustment for the active engineering task.

Goal: Capture a small-to-medium course correction that the user realized mid-task (not from external review), record it as a decision in DECISIONS.md, update TASK_STATE.md to reflect the adjusted direction, and route back to the appropriate command to continue the work.

Mandatory context bootstrap (before any output):

• Read these sections in WORKFLOW_OPERATING_SYSTEM.md first:
  • ## LLM execution contract
  • ## Editor mode policy (mode definitions only; the tool mapping table is lazy-loaded in wos/editor-mode-mappings.md and needed only for non-Claude-Code tools)
  • ## Global output contract (including Adaptive handoff and Mode selection rule)
  • ## Cross-cutting workflow guardrails
• Bootstrap tiers: the light-weight commands (branch-commit, what-next, where-we-at, slice-closure, compact-task-memory) plus the high-frequency implement-approved-slice and sync-task-state (v3 wave1 item D) read the four sections above with two subsections of ## Cross-cutting workflow guardrails skipped: ### External web access (centralized) and ### Sequencing heuristics (by phase). Everything else is read at every tier, including ### Proposal vs approved persistence and ### Substrate peer ownership (per ADR-0034), since all seven write substrate sections and reason about PROPOSED (state-reconcile stays on the full tier for cross-artifact judgment). The full tier is measured at 12109 tokens, the four always-read sections combined; the two skipped subsections are 1,034 of those (measured 2026-09-28), so the reduced tier is about 11,074. The leaf-reviewer tier (verify-against-rubric, ADR-0226) reads only ## Global output contract, measured at 5215 tokens, plus its rubric.
• Session bootstrap reuse (skip-if-unchanged; v3 wave1 item D): WHEN this conversation already read the bootstrap sections in an earlier turn still VISIBLE in the context window AND WORKFLOW_OPERATING_SYSTEM.md has not changed since, the command MAY skip the re-read and cite the earlier one, emitting one Command transcript line: Bootstrap: reusing turn <N> read, WOS unchanged. Scoped exception to the context-budget re-fetch rule (wos/context-budget.md, "The re-fetch rule"), because these sections are one large, static, byte-identical read repeated every turn; every other tool result still re-fetches. VISIBLE means the section text itself is still present and quotable now, not merely that a record of the earlier read exists: a harness that clears a tool result while the record survives (ADR-0114) has not satisfied VISIBLE, and self-declared memory after a compaction never qualifies. A stateless-per-turn harness is excluded. The transcript line is mandatory; a silent skip is invalid 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 additional sections only when relevant to this command's role.
• Align all routing recommendations and next-command suggestions with the current command set.
• Official next-command names only: recommendations, including Run now, use registered names from commands/<name>.md or commands/<name>/SKILL.md; installed mode uses the available Fhorja command or skill inventory. No invented aliases. Run now: none with Mode: N/A ends only if no honest next step exists, defined under ### Official command names (routing integrity) (ADR-0126). Required inputs:
• active task folder path
• TASK_STATE.md (must reflect current phase and last completed step)
• DECISIONS.md (current canonical decisions)
• IMPLEMENTATION_PLAN.md (current plan that is being adjusted)
• the user's description of the realization: what was being done, what was noticed, what should change
• optional: which slice or phase the adjustment affects

Task repository files to update:

• DECISIONS.md (append a new decision recording the adjustment with a D-N: mid-task adjustment prefix, or a provisional ### P-N under ## Provisional decisions per the attended-chain rule below)
• DECISIONS.md ## Decision history (WHEN the mid-task realization is new evidence that contradicts an already-persisted NON-decision claim, record a defeasible-claim revision here per the ADR-0109 / D-10 write rule in wos/substrate-peers.md ## Decision history: append-only, name the contradicting evidence and its provenance rank, mark [OPEN]; task-close blocks on an unresolved one)
• TASK_STATE.md (update Last completed step if it is now wrong, update Recommended next step to reflect the adjusted direction, update Risks to watch if the adjustment introduces new risk)
• IMPLEMENTATION_PLAN.md (only if the adjustment changes plan text; default is to leave plan as-is and note the adjustment in DECISIONS.md)
• relevant SLICES/*.md (only if the active slice scope must change to reflect the adjustment)

Operating rules:

• Do not implement production code.
• Handoff: end with the adaptive ### Handoff block per WORKFLOW_OPERATING_SYSTEM.md ## Global output contract (Mode A compact or Mode B full).
• Substrate write protocol (per ADR-0034, K.2). MANDATORY for every substrate section this command writes (the DECISIONS.md section receiving the new D-N entry, which is ## Locked decisions, plus ## Decision history when this turn records a defeasible-claim revision, and each updated TASK_STATE.md section). Follow commands/_shared/substrate-write-protocol.md: emit the <!-- wos:write owner=direction-adjust section='## X' ... --> transaction header and append one JSONL line per section write to active/<task>/.wos/VERIFICATION_LOG.jsonl. Header placement per commands/_shared/substrate-write-protocol.md ## Transaction header: the transaction header goes on its own line IMMEDIATELY above the ## <section> heading line. NEVER place it below the heading, and NEVER above a ### D-N entry inside the section; a header placed below the heading leaves the section counted as header-less by the K.4 drift scan (scripts/scan-substrate-headers.sh), which checks only the line immediately preceding each ## heading.
• Treat the adjustment as a new decision, not a silent edit. The output must produce a numbered entry (D-N, or P-N under the attended-chain rule below) for DECISIONS.md.
• Make the adjustment auditable: clearly show what direction was being followed before, what changed, and why.
• Validate the adjustment against existing locked decisions and invariants. If the adjustment contradicts a locked decision, surface this explicitly and route to decision-interview instead of silently overwriting.
• Validate the adjustment against INVARIANTS_AND_NON_GOALS.md if present. If the adjustment crosses a non-goal or invalidates an invariant, surface this and require user confirmation before proceeding.
• Attended chain on a task branch (ADR-0233; the P-N format rests on its provisional P-1). WHILE this command runs in an attended chain on a task branch (a Task branch: line in TASK_STATE.md ## Resume notes), no Operating mode: assisted is declared, and the realization is the agent's rather than one the person described, it SHALL be recorded as a provisional ### P-N under DECISIONS.md ## Provisional decisions, in the format commands/decision-interview.md gives its provisional mode, instead of a D-N: mid-task adjustment, and the chain SHALL continue. The two conflict rules above do not wait for a person in such a chain. IF the adjustment contradicts a locked D-N or crosses an invariant or non-goal THEN this command SHALL NOT apply it: the work keeps following the locked decision or the invariant, the adjustment is recorded as a P-N with Impact: high naming the D-N or the invariant it would change, and the chain continues on the current direction until the person answers. An unattended or fleet-dispatched run (ADR-0237) keeps the rules above; the one exception is a session under standing authorization (ADR-0244, wos/cross-cutting-workflow-guardrails.md ### Unattended sessions), which SHALL follow this bullet.
• Keep the new DECISIONS.md entry concise: 2-5 lines covering what changed and why.
• Update TASK_STATE.md minimally: only fields actually affected by the adjustment. Do not rewrite unrelated sections.
• If the adjustment requires re-planning (the slice scope is now wrong, or the slice order needs to change), the recommended next command must be implementation-plan or state-reconcile, not blind continuation.
• If the adjustment is small enough that the current slice can absorb it without re-planning, the recommended next command is the slice-execution command appropriate to the phase (typically implement-approved-slice).
• Treat task-memory write policy per WORKFLOW_OPERATING_SYSTEM.md: write the file and mark it APPLIED.

Required output:

1. Summary of the adjustment in plain language: "before, the direction was X; now, the direction is Y; the trigger was Z".
2. The new DECISIONS.md entry to append, formatted as D-N: mid-task adjustment - <one-line title> followed by 2-5 lines of detail.
3. The updated fields in TASK_STATE.md (only the fields that actually change).
4. The validation result against locked decisions: compatible, requires decision-interview, violates invariant: <which one>, or, in an attended chain on a task branch, recorded as provisional P-N (with not applied when it conflicts).
5. The recommended next command, with reasoning about whether the adjustment requires re-planning or fits the current slice.
6. Recommended editor mode for that next command.
7. Recommended work complexity (LOW | MEDIUM | HIGH | N/A) for the next step.

Substrate digest fallback

Digest fallback when the canonical helper is unreachable. sha_of_section extracts a section's body with awk and pipes it to shasum -a 256. A run inside a permission boundary that admits shasum but refuses awk and sed cannot invoke that helper, and MUST NOT reimplement it. Assume the helper is unreachable whenever the workflow repository's scripts/ directory is not readable from the working directory.

WHERE the canonical per-section digest helper is unreachable, the write SHALL use a whole-file SHA-256 and SHALL declare the reduced scope:

# One call per file. Read the hash (the first field) out of the output;
# do not pipe it through `cut` or `awk` to trim it.
shasum -a 256 TASK_STATE.md

Then add "sha_scope":"file" to that JSONL line, so a validator distinguishes a file digest from a section digest. The rest of the protocol is unchanged: the transaction header still goes above the section heading, and there is exactly one JSONL line per section write.

Do NOT rebuild the helper by writing each section out as its own file to hash it separately.

At file scope two sections written in the same run share a digest, so the sha_scope field is mandatory, not optional.

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

• List files in the task repository that would change, or None.
• For each file, mark APPLIED / PROPOSED / SKIP and follow the task-memory write policy in WORKFLOW_OPERATING_SYSTEM.md (task memory is APPLIED in every mode; a section this command does not own gets a <!-- PROPOSED by <command>: ... --> block for its owner, ADR-0034).
• Default for this command: APPLIED patches on DECISIONS.md and TASK_STATE.md; conditionally on IMPLEMENTATION_PLAN.md or SLICES/*.md only when the adjustment requires it.

Command transcript

• Keep this section operational and brief; do not restate file content already listed in ### Artifact changes.
• Max 4 lines in normal runs.
• Max 3 lines in no-op runs (including NO_OP_TRACE).
• Include NO_OP_TRACE (1-3 lines) if the realization is too small to record as a decision (route to capture-observation instead) or if the adjustment turns out to require a heavier path (route to decision-interview or state-reconcile).

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 adjustment is recorded as a numbered D-N: mid-task adjustment entry in DECISIONS.md with concise reasoning, or, when it is the agent's own realization in an attended chain on a task branch, as a provisional ### P-N with its Evidence:, Impact: and Status: provisional lines.
• Every substrate write carries its wos:write transaction header on its own line IMMEDIATELY above the ## <section> heading, never below the heading and never above a ### D-N entry inside the section, per commands/_shared/substrate-write-protocol.md ## Transaction header.
• TASK_STATE.md updates are minimal and target only fields affected by the adjustment.
• Validation against locked decisions and invariants is explicit; the output names any conflict and routes to a heavier command rather than silently overwriting.
• The adjustment never overrides locked decisions in place. Conflicts route to decision-interview, or, in an attended chain on a task branch, are recorded as an Impact: high P-N while the work keeps following the locked decision.
• The adjustment never violates invariants or crosses non-goals silently. Conflicts surface for user confirmation; in an attended chain on a task branch that confirmation is asked in the draft pull request through the P-N, not by stopping.
• Artifact changes marks each patch as APPLIED.
• Handoff has non-empty Run now, Mode, Work complexity, and Reason fields, plus Resume context only in Mode B; ending after the decision entry without a Handoff is invalid output.
• The recommended next command is appropriate to the size of adjustment: small adjustments resume current slice; medium adjustments trigger implementation-plan or state-reconcile; conflicts trigger decision-interview.
• 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 auditability of the direction change, protection of locked decisions and invariants, minimal disruption to in-progress work, and clear routing to the right next step based on the size of the adjustment.