coilyco-flight-deck/ward

Fork 0

polish batch, run 16: agent-legibility #465

New issue

Closed

opened 2026-07-01 22:32:11 +00:00 by coilyco-ops · 2 comments

coilyco-ops commented

2026-07-01 22:32:11 +00:00

Member

Why

Non-gating friction from run 16 (the-agent-itself / agent-legibility), batched per sweep rules:

docs/agent-preflight.md, step "1." of the numbered pre-flight sequence, is a single ~350-word run-on paragraph covering at least five distinct behaviors with no sub-structure: prompt construction, comment-thread weighting (and the fact wards own automated comments are stripped from that thread - a non-obvious, load-bearing detail buried mid-paragraph), one-shot execution mechanics (claude -p vs goose run -t), and the neutral-temp-dir rationale (ward#169). A harness (or human) searching for one fact - e.g. "does the pre-flight read run in the dispatch cwd or a fresh clone?" - has to parse the entire block; nothing below the numbered-list level is addressable by heading or bullet. Breaking this into labeled sub-bullets (prompt shape / thread filtering / execution mechanics / cwd isolation) would cut the context cost of finding any one fact without losing content.

Deliverable / Done condition

Checkbox independently fixable; done when checked.

Severity: polish · persona: the-agent-itself · angle: agent-legibility
Filed by Claude Code during a cold-read release pressure test (run 16).

## Why Non-gating friction from run 16 (`the-agent-itself` / `agent-legibility`), batched per sweep rules: - [ ] `docs/agent-preflight.md`, step "1." of the numbered pre-flight sequence, is a single ~350-word run-on paragraph covering at least five distinct behaviors with no sub-structure: prompt construction, comment-thread weighting (and the fact wards own automated comments are stripped from that thread - a non-obvious, load-bearing detail buried mid-paragraph), one-shot execution mechanics (`claude -p` vs `goose run -t`), and the neutral-temp-dir rationale (ward#169). A harness (or human) searching for one fact - e.g. "does the pre-flight read run in the dispatch cwd or a fresh clone?" - has to parse the entire block; nothing below the numbered-list level is addressable by heading or bullet. Breaking this into labeled sub-bullets (prompt shape / thread filtering / execution mechanics / cwd isolation) would cut the context cost of finding any one fact without losing content. ## Deliverable / Done condition Checkbox independently fixable; done when checked. --- **Severity: polish** · persona: the-agent-itself · angle: agent-legibility Filed by Claude Code during a cold-read release pressure test (run 16).

coilyco-ops added the

pressure-test

label

2026-07-01 22:32:11 +00:00

coilyco-ops referenced this issue

2026-07-01 22:33:07 +00:00

meta: release pressure-test run log (cold-read usability sweeps) #428

coilyco-ops added this to the ward launch milestone

2026-07-01 23:29:31 +00:00

coilyco-ops added the

headless

labels

2026-07-01 23:30:25 +00:00

coilyco-ops referenced this issue

2026-07-02 00:01:42 +00:00

meta: jul 1st evening sprint - launch burndown session 1 #492

coilyco-ops commented

2026-07-02 07:10:03 +00:00

Author

Member

🔒 Reserved by ward agent --driver claude — container engineer-claude-ward-465 on host kais-macbook-pro-2.local is carrying this issue (reserved 2026-07-02T07:10:00Z). Concurrent ward agent runs are blocked until it finishes or the reservation goes stale (2h0m0s TTL); --force overrides.

— Claude (she/her), via ward agent

🔒 Reserved by `ward agent --driver claude` — container `engineer-claude-ward-465` on host `kais-macbook-pro-2.local` is carrying this issue (reserved 2026-07-02T07:10:00Z). Concurrent `ward agent` runs are blocked until it finishes or the reservation goes stale (2h0m0s TTL); `--force` overrides.  — Claude (she/her), via `ward agent`

coilysiren referenced this issue

2026-07-02 07:10:55 +00:00

meta: jul 1st evening sprint - launch burndown session 1 #492

coilysiren referenced this issue from a commit

2026-07-02 07:13:34 +00:00

docs(agent-preflight): sub-structure the pre-flight step 1 (closes #465)

coilysiren closed this issue

2026-07-02 07:13:34 +00:00

coilysiren referenced this issue from a commit

2026-07-02 07:13:34 +00:00

Merge issue-465: sub-structure agent-preflight step 1

coilysiren commented

2026-07-02 07:13:53 +00:00

Owner

WARD-OUTCOME: done - split agent-preflight.md step 1 into four labeled sub-bullets (prompt shape / thread filtering / execution mechanics / cwd isolation).

Genuinely small change, but it fought back in one spot: the doc was already at 3991 chars against a hard 4000-char cap, and adding four bullet labels + list structure blew past it to 4252. So most of the work was not the restructure - it was tightening the prose back down under the cap without dropping a single fact or ward# ref. Landed at 3982, which leaves almost no headroom, so the real fragility here is that agent-preflight.md is a dense doc bumping its size ceiling.

Confident in the result: the four facts the issue called out (especially the load-bearing "ward's own automated comments are stripped from the thread" detail) are now each reachable by label, and the full pre-commit suite passes. Rough edge / possible follow-up: this doc is one edit away from the cap again, so the next substantive addition to the pre-flight sequence will likely force a split into a child docs/*.md. Worth filing if/when that pressure returns.

WARD-OUTCOME: done - split agent-preflight.md step 1 into four labeled sub-bullets (prompt shape / thread filtering / execution mechanics / cwd isolation). Genuinely small change, but it fought back in one spot: the doc was already at 3991 chars against a hard 4000-char cap, and adding four bullet labels + list structure blew past it to 4252. So most of the work was not the restructure - it was tightening the prose back down under the cap without dropping a single fact or ward# ref. Landed at 3982, which leaves almost no headroom, so the real fragility here is that agent-preflight.md is a dense doc bumping its size ceiling. Confident in the result: the four facts the issue called out (especially the load-bearing "ward's own automated comments are stripped from the thread" detail) are now each reachable by label, and the full pre-commit suite passes. Rough edge / possible follow-up: this doc is one edit away from the cap again, so the next substantive addition to the pre-flight sequence will likely force a split into a child docs/*.md. Worth filing if/when that pressure returns.

coilysiren referenced this issue

2026-07-02 07:33:03 +00:00

Salvage should post back to the carried issue that spawned the run, and must not salvage runs whose work already landed #518

coilysiren referenced this issue

2026-07-14 15:27:09 +00:00

director should emit colorized, space buffered, visually distinct messages about the categories of actions its performing on startup #1283