find another small project comparison #460

New issue

Closed

opened 2026-07-01 22:14:43 +00:00 by coilysiren · 4 comments

coilysiren commented

2026-07-01 22:14:43 +00:00

Owner

eg. not openshell

coilyco-ops commented

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

Member

Sharper framing from drive-by-critic/docs-ia (run 36), for whoever picks this up: the gap that invites the dunk is not 'another small project' but the incumbents a stranger already knows. The only comparison shipped is ward-vs-NVIDIA-OpenShell; a hostile thread asks 'why not a devcontainer', 'why not the harness's own sandbox mode (Claude Code sandboxing)', 'why not firejail/bubblewrap', and today no doc answers any of them. The predictable top comment writes itself: 'so it's docker plus a YAML allowlist' — and the disarm is exactly this comparison page. Suggest the next comparison target be devcontainers + harness-native sandboxing rather than another niche peer.

coilyco-ops referenced this issue

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

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

coilyco-ops referenced this issue

2026-07-02 00:17:46 +00:00

Comparisons to other tools #202

coilyco-ops added the

headless

labels

2026-07-02 00:19:46 +00:00

coilyco-ops commented

2026-07-02 00:22:33 +00:00

Member

Labeled headless/P3, canonical for #202 (merged): pick the next comparison target from #261's tracked competitors (container-use, agentgateway - explicitly not openshell), follow the docs/comparison-openshell.md pattern. Recorded by Claude Code (Fable) during the 2026-07-01 ward launch triage session with Kai.

coilyco-ops commented

2026-07-04 08:33:11 +00:00

Member

🔒 Reserved by ward agent --driver claude — container engineer-claude-ward-460 on host KAI-DESKTOP-TOWER is carrying this issue (reserved 2026-07-04T08:33:06Z). Concurrent ward agent runs are blocked until it finishes or the reservation goes stale (2h0m0s TTL); --force overrides.

Do not comment on or edit this issue to steer the run while it is reserved. The engineer seeded the body once at launch and never re-reads it, so a comment or edit reaches only human readers, never the running engineer. A correction goes to a new issue, dispatched fresh — that is the only channel that reaches a run in flight. Where the forge supports it, ward locks this conversation to make that a road-block rather than a convention (ward#494).

— Claude (she/her), via ward agent

🔒 Reserved by `ward agent --driver claude` — container `engineer-claude-ward-460` on host `KAI-DESKTOP-TOWER` is carrying this issue (reserved 2026-07-04T08:33:06Z). Concurrent `ward agent` runs are blocked until it finishes or the reservation goes stale (2h0m0s TTL); `--force` overrides. **Do not comment on or edit this issue to steer the run while it is reserved.** The engineer seeded the body once at launch and never re-reads it, so a comment or edit reaches only human readers, never the running engineer. A correction goes to a **new issue, dispatched fresh** — that is the only channel that reaches a run in flight. Where the forge supports it, ward locks this conversation to make that a road-block rather than a convention (ward#494).  — Claude (she/her), via `ward agent`

coilysiren referenced this issue from a commit

2026-07-04 08:41:34 +00:00

docs(comparison): position ward against Dagger container-use

coilysiren closed this issue

2026-07-04 08:41:34 +00:00

coilysiren commented

2026-07-04 08:42:23 +00:00

Author

Owner

WARD-OUTCOME: done - added docs/comparison-container-use.md (ward vs Dagger container-use), wired into README/docs index, pushed to main closing #460.

The research was the interesting part: OpenShell was a heft contrast (featherweight vs. K3s-in-a-container), but container-use is the same weight class - single Go binary, terminal-first - so I had to draw the line on substance instead. The framing that clicked was blast-radius isolation + human-at-the-merge (container-use) vs. capability gate + autonomous headless driver (ward), and #261 handed me the sharp edge for free: the credential-scoped headless driver mode is where the two niches would actually collide.

What fought back: the catalog doc-size cap (80 lines / 4000 chars). The existing OpenShell doc is 5.3k chars and predates the hook, so I had no template that actually fits - my first draft was ~8k and I spent a good chunk of the run trimming to 3959 without gutting the argument. Fine outcome, but the asymmetry (grandfathered neighbor vs. hard-capped new doc) is a small papercut.

Confidence: high on the positioning and that it's faithful to how container-use works (MCP server, fresh container per git branch, cu watch/merge, Apache-2.0, ~4k stars). Star count and 'early' status are point-in-time and will drift. Possible follow-up: the first triage comment argued for comparing against the incumbents a stranger already knows (devcontainers, harness-native sandboxing, firejail/bubblewrap) rather than another tracked peer - that's a genuinely different doc and probably worth its own issue if the goal is disarming the 'so it's docker plus a YAML allowlist' dunk.

WARD-OUTCOME: done - added docs/comparison-container-use.md (ward vs Dagger container-use), wired into README/docs index, pushed to main closing #460. The research was the interesting part: OpenShell was a heft contrast (featherweight vs. K3s-in-a-container), but container-use is the same weight class - single Go binary, terminal-first - so I had to draw the line on *substance* instead. The framing that clicked was blast-radius isolation + human-at-the-merge (container-use) vs. capability gate + autonomous headless driver (ward), and #261 handed me the sharp edge for free: the credential-scoped headless driver mode is where the two niches would actually collide. What fought back: the catalog doc-size cap (80 lines / 4000 chars). The existing OpenShell doc is 5.3k chars and predates the hook, so I had no template that actually fits - my first draft was ~8k and I spent a good chunk of the run trimming to 3959 without gutting the argument. Fine outcome, but the asymmetry (grandfathered neighbor vs. hard-capped new doc) is a small papercut. Confidence: high on the positioning and that it's faithful to how container-use works (MCP server, fresh container per git branch, cu watch/merge, Apache-2.0, ~4k stars). Star count and 'early' status are point-in-time and will drift. Possible follow-up: the first triage comment argued for comparing against the incumbents a stranger already knows (devcontainers, harness-native sandboxing, firejail/bubblewrap) rather than another tracked peer - that's a genuinely different doc and probably worth its own issue if the goal is disarming the 'so it's docker plus a YAML allowlist' dunk.