docs(stories): pin tasks install auto-detect algorithm in story 03#62
Merged
Conversation
closes user-stories-install-detect-docs Story 03's first line sold `tasks install` as "auto-detects agents" but never named which directories triggered which install, nor what happened when the dir was missing. The "Or copy manually:" table that followed was a workaround, not documentation of the auto-detect behavior. Story 03 now has an "Auto-detect algorithm" section with a four-column table (Agent / Detection signal / Installed path / Source file) sourced directly from `AGENT_MAPPINGS` in `packages/cli/src/commands/install.ts`: | Agent | Detection | Installed | |-------------|---------------|--------------------------------------| | Claude Code | `.claude/` | `.claude/skills/next-task/` | | Codex | `.agents/` | `.agents/skills/next-task/` | | Cursor | `.cursor/` | `.cursor/commands/next-task.md` | | Devin | `.devin/` | `.devin/skills/next-task/` | | Gemini CLI | `.gemini/` | `.gemini/commands/next-task.toml` | | Windsurf | `.windsurf/` | `.windsurf/workflows/next-task.md` | Plus the fallback contract: when the detection signal is missing and `--all` was not passed, the agent is skipped silently and `tasks install` exits 0 even if zero agents matched. The `docs/user-stories/README.md` "How" column for story 3 now deep-links to the new section and names all 6 agents the algorithm covers. A new test in `install.test.ts` (`pins the AGENT_MAPPINGS install paths for all 6 agents`) asserts every destination path the doc table claims, so changing a path in `install.ts` without updating the story (or vice versa) fails CI. Verified: npm run build, npm test (404 tests, +1 new), npm run lint, npx -y @tasks-md/lint TASKS.md — all clean. --- _🤖 Written by an agent, not Fyodor. Ping me if this looks off._
fyodoriv
added a commit
that referenced
this pull request
May 3, 2026
…#68) closes user-stories-crosslink-audit After the recent additions (story 08, runnable demos, install table, worked example) several stories reference each other. Audit walked every Markdown link in docs/user-stories/, README.md, packages/cli/ README.md, packages/mcp/README.md, and the package READMEs and verified each path + anchor against the live tree. Findings: Working: every cross-story link in docs/user-stories/ + README.md resolves to an existing file and a real heading slug. The Auto-detect anchor cross-link added in PR #62 still works; the story 06/07 anchor links in user-stories/README.md still work. Broken paths fixed: packages/mcp/README.md:81 [MIT](../LICENSE) -> ../../LICENSE packages/lint/README.md:102 [MIT](../LICENSE) -> ../../LICENSE Both pointed at packages/LICENSE which doesn't exist; LICENSE lives at the repo root, two dirs up. Inverse / back-links added: Story 02 metadata table now links **ID** + **Blocked by** to Story 04 (the source of truth for blocker ordering) and **Tags** to Story 05 #tag-based-routing (the routing contract). The table footer points at Story 08 for the agent-managed fields (**Blocked**, **Research**, **Last-enriched**, **Parent**) and project-level policies — closing the loop that those fields are defined in detail elsewhere. Story 04 #2 ("Tag preference") was a plain-prose "see Story 05" sentence; now a Markdown hyperlink to Story 05 #tag-based- routing so a reader can jump to the routing contract directly. Verified by a Python anchor-checker that walks every Markdown link in the audited files and asserts: - the path component resolves to an existing file - the anchor matches a heading slug or {#explicit-anchor} in the target Output: "All cross-links resolve cleanly!" on the post-fix tree. Verified: npm run lint, npx -y @tasks-md/lint TASKS.md — 0 errors. --- _🤖 Written by an agent, not Fyodor. Ping me if this looks off._
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.
Why this is needed
Closes `user-stories-install-detect-docs` (P2 hardening from PR #58).
Story 03's first "Install" section sold `tasks install` as "auto-detects agents" but never named which directories triggered which install, nor what happened when the agent dir was missing. The "Or copy manually:" table that followed was a workaround, not documentation of the actual algorithm — readers had to read `install.ts` to find out what `tasks install` would actually do in their repo.
Hardening the story means making the contract reader-verifiable.
Summary
Story 03 now has an Auto-detect algorithm section with a four-column table — Agent / Detection signal / Installed path / Source file — sourced directly from `AGENT_MAPPINGS` in `install.ts`:
Plus the fallback contract: when the detection signal is missing and `--all` was not passed, the agent is skipped silently and `tasks install` exits 0 even if zero agents matched. Use `--all` to scaffold every agent's directory regardless, or `--agent ` to target a single entry.
The `docs/user-stories/README.md` "How" column for story 3 now deep-links to the new section and names all 6 agents the algorithm covers.
Pin: regression guard
A new test in `install.test.ts` — `pins the AGENT_MAPPINGS install paths for all 6 agents` — asserts every destination path the doc table claims (`.claude/skills/next-task/SKILL.md`, `.agents/skills/next-task/SKILL.md`, `.cursor/commands/next-task.md`, `.devin/skills/next-task/SKILL.md`, `.gemini/commands/next-task.toml`, `.windsurf/workflows/next-task.md`). Changing a destination path in `install.ts` without updating the story (or vice versa) fails CI.
Test plan
🤖 Written by an agent, not Fyodor. Ping me if this looks off.