docs: refresh post-gantry-cutover; drop stale Claude Code framing - #50
Open
barryroodt wants to merge 1 commit into
Open
docs: refresh post-gantry-cutover; drop stale Claude Code framing#50barryroodt wants to merge 1 commit into
barryroodt wants to merge 1 commit into
Conversation
The README still described wrily as 'using Claude Code'. Post-cutover the runtime is the gantry subprocess and the tool is provider-agnostic (Anthropic / OpenAI / Google / OpenRouter). Swept all live docs for the same drift. - README: headline rewritten (gantry harness, provider-agnostic); 'Claude's skill set' -> 'reviewer's skill set'; 'Claude reads CLAUDE.md naturally' -> 'Wrily's reviewer reads CLAUDE.md / AGENTS.md'; fixed WRILY_AGENT_TIMEOUT_MS default (30 min -> 12 min, matches DEFAULT_TIMEOUT_MS in agent/gantry.ts). - self-hosting: provider matrix now includes openrouter; 'Anthropic API key' / 'Claude Code OAuth token' framing generalized to provider-agnostic. - adoption: provider-agnostic prose; cost FAQ no longer presents hard USD figures as universal (model/provider-dependent; free with OpenRouter free models). - webhook-architecture: 'Claude timeout' / 'Claude Code session' -> agent / gantry subprocess run. - followups: max_budget_usd -> max_tokens; pi-cutover section flagged as superseded by the gantry cutover (#40). - writing-skills: skill discovery attributed to the reviewer, not the (stale) agent-team-review skill. Kept intentionally: .claude/skills/ + CLAUDE.md (live conventions the prompt reads), the 'anthropic' env name, CLAUDE_CODE_OAUTH_TOKEN (real Anthropic auth), current model slugs (claude-opus-4-8, claude-3.5-sonnet), and historical changelog entries. Out of scope (flagged): skills/agent-team-review/SKILL.md still describes Claude Code Agent Teams (TeamCreate / CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS); gantry now owns team orchestration via profiles/review/{compose,subagent,unify}.md and system.md already guards against the skill's CI/markdown mandates. Rewriting it is a behavioral change needing team-mode verification — separate from this docs pass.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
What
The README still opened with "AI-powered code review agent using Claude Code" — stale since the gantry cutover (#40) and the OpenRouter add (#47). Wrily now drives the in-house gantry subprocess and is provider-agnostic (Anthropic / OpenAI / Google / OpenRouter). I swept every live doc for the same drift.
Changes (6 files, +21/-21)
Claude's skill set->the reviewer's skill set; "Claude readsCLAUDE.mdnaturally" -> "Wrily's reviewer readsCLAUDE.md/AGENTS.md"; fixedWRILY_AGENT_TIMEOUT_MSdefault: 30 min -> 12 min (matchesDEFAULT_TIMEOUT_MS = 12 * 60 * 1000insrc/agent/gantry.ts).max_budget_usd->max_tokens; the pi-cutover section is flagged as superseded by the gantry cutover (Replace pi-coding-agent with gantry subprocess (cutover) #40) (PiRunnerno longer exists).agent-team-reviewskill.Kept intentionally (verified accurate, not drift)
.claude/skills/+CLAUDE.md/AGENTS.md— live conventions the prompt template literally reads (src/prompt/templates.ts,instructions.ts, gantrysystem.md).anthropicGitHub Actions environment name (an identifier),CLAUDE_CODE_OAUTH_TOKEN(a real Anthropic auth method), and current model slugs (claude-opus-4-8,claude-3.5-sonnet).Out of scope — flagged for a separate PR
skills/agent-team-review/SKILL.mdstill documents Claude Code Agent Teams (TeamCreate/SendMessage/CLAUDE_CODE_EXPERIMENTAL_AGENT_TEAMS=1). Post-cutover, gantry owns team orchestration viaprofiles/review/{compose,subagent,unify}.md, andprofiles/review/system.mdalready contains explicit guards overriding that skill's "Run CI" / markdown-output mandates. Rewriting (or removing) the skill is a behavioral change that needs team-mode verification against gantry — not a prose-docs edit — so it's deliberately excluded here.Verification
Pure documentation; no code touched. Grep sweep confirms no remaining stale
Claude Code/pi/max_budget_usd/ Cursor / retired-slug references in live prose.