repo-init-manual-fallback
Use inside repo-init-scan after the official stage whenever scaffold verification, bounded repair, or `best-copilot.md` rewrite is still needed.
Skill by funky-eyes
Chalk Labs
Need help deploying repo-init-manual-fallback?
Chalk Labs sets up agents like this inside real teams: connected to your tools, tested on your data, with a human in the loop where it matters.
Install
Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.
npx agentmarket install skill/funky-eyes-repo-init-manual-fallback
Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.
Get the weekly diff
The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.
About
Reviews
Write a reviewNo reviews yet. Installed it? Reviews come from verified installers only. How reviews work
Permissions
What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.
- Tools
- No tools found
- Secrets
- None asked for
- Install scripts
- None found
- Runs as
- Shell scripts
Network hosts it talks to 1
codeload.github.com· found in install command
Files it touches
~/.claude/skills/repo-init-manual-fallback
What's inside
- Skill name
repo-init-manual-fallback- Description
- Use inside repo-init-scan after the official stage whenever scaffold verification, bounded repair, or `best-copilot.md` rewrite is still needed.
- License
- Apache-2.0
- Instructions
- 96 lines in SKILL.md · view on GitHub
Before installing
- Ships executable scripts
Files (3)
- SKILL.md
- inline-templates-reference.md
- bootstrap-target-scaffold.sh · scripts/
SKILL.md (first lines)
# Repo Init Manual Fallback Use this stage inside `repo-init-scan` whenever scaffold verification, bounded repair, or `best-copilot.md` rewrite is still needed. After official init succeeds, this stage still owns the final scaffold verification barrier and sentinel rewrite. ## Immediate Execution Rule When this stage has just been loaded inside `repo-init-scan`, the next assistant action must create/repair/verify the required init artifacts or return `BLOCKED`, then emit `## Repo Init Manual Fallback`. Do not stop at `Skill(...repo-init-manual-fallback...) Successfully loaded`, and do not inspect business source except the bounded evidence allowed below. ## Boundary - This stage is fail-closed. Do not continue to requirements analysis, planning, implementation, or security rewrites until the required files have been created or a `BLOCKED` result has been returned. - Reading package/build files is allowed only as bounded evidence for creating `.github/instructions/project.instructions.md`. - Write persistent state into the target repository, never into the plugin installation/cache directory. - Loading this skill or the bootstrap skills is not enough. Success requires actual create/repair operations followed by path and content checks on disk. - In shell-capable runtimes, first run the bundled deterministic helper when its path is discoverable. If it cannot be located, perform the documented create-or-repair fallback inline instead of blocking. - If `target-instructions-bootstrap`, `target-memory-bootstrap`, or `target-spec-bootstrap` cannot be invoked mechanically, read their `SKILL.md` templates and perform the documented create-only repair inline. - Do not hand-write improvised versions of scaffold files. When creating files inline (no shell helper), use the exact template content from `inline-templates-reference.md`. When a template is not there, read the corresponding bootstrap skill for the full template before writing. - The `best-copilot.md` sentinel must be written ONLY after all runtime-required artifacts exist and pass content verification. - **PM inline analysis requirement**: When creating `.github/instructions/project.instructions.md` inline (no shell helper), the PM MUST analyze the target repository before writing. Do NOT write "unknown" for fields that can be determined by reading the repository: 1. Read the root build file (`pom.xml`, `build.gradle`, `package.json`, `go.mod`, etc.) for build system, modules, dependencies. 2. For Maven: parse `<modules>` from root pom.xml, identify main module, find `@SpringBootApplication` or `main()`. 3. Search for security/auth files (`*Security*`, `*Auth*`, `*OAuth*`, `WebSecurityConfig*`). 4. Read `application.yml`, `application.properties`, Docker files for runtime requirements. 5. Infer build/test commands from build file and wrapper scripts (`mvnw`, `gradlew`). 6. Record genuine `unknown` only for things undetectable without running the application. - A `project.instructions.md` with more than 3 "unknown" values for detectable fields is a protocol violation. - **PM post-script enhancement**: When the shell helper creates `.github/instructions/project.instructions.md` but the file contains "unknown" for detectable fields, the PM MUST read the file, analyze the target repository, and update those fields with real data. Do not accept shell-script "unknown" for fields determinable by reading source files. ## Sentinel Ownership This skill owns constructing target-root `best-copilot.md`. The companion `repo-init-gate` skill owns validating it, and `repo-init-scan` owns orchestrating when this stage may write it. ## Required First-Use Artifacts Before `next_task_ready: yes`, verify these paths in the target repository: - `.github/instructions/project.instructions.md` - `.github/instructions/must.instructions.md` - `.github/instructions/skills-index.instructions.md` - `CLAUDE.md` (when Claude Code compatibility is required) - `AGENTS.md` (when Codex compatibility is required) - `.codex/agents/senior-project-expert.toml`, `.codex/agents/technical-architect.toml`, `.codex/agents/developer.toml`, `.codex/agents/frontend-designer.toml`, `.codex/agents/quality-assurance-expert.toml`, `.codex/agents/security-reviewer.toml`, `.codex/agents/specification-writer.toml`, `.codex/agents/root-cause-fixer.toml` (when Codex compatibility is required) - `memories/README.md`, `memories/repo/INDEX.md`, `memories/repo/current-workstreams.md`, `memories/repo/project-state.md`, `memories/repo/workflow-rules.md`, `memories/repo/decisions.md`, `memories/repo/logs/README.md`, `memories/repo/archive/deprecated-decisions.md` - `spec/INDEX.md`, `spec/templates/requirements-template.md`, `spec/templates/design-template.md`, `spec/templates/tasks-template.md` `best-copilot.md` is the verified-init sentinel for contract version `0.8.0`. Must be written only after the other runtime-required artifacts and content checks pass. The canonical write content is the exact 3-line YAML frontmatter block -- no headings, descriptions, or dates. Repeat-request readers use only the frontmatter version for the cheap skip decision. ## Required Content Checks - `must.instructions.md` contains: `## Request Flow`, `## Per-Request Hard Gates`, `### PM Native Ask Trigger Gate`, `## Shared State Contracts`, `## Search Precision`, `## Command Output Budget`, `## Memory And Spec`, `## Agents and Dispatch`, `## Implementation and Verification`, plus `NEEDS_USER_INPUT`, `BLOCKED`, `work_mode`, `task_type`, `pm_action`, `fixed-string-before-regex`, `vscode_askQuestions`, `Self-evolution`, and `evolution_signal`. - `skills-index.instructions.md` contains bootstrap routing and `## Claude Code Skill Names`. - `CLAUDE.md` references all three `.github/instructions/` files (when Claude Code compatibility is required). - `AGENTS.md` references `.codex/agents/*.toml`, and every Codex TOML adapter has the expected `name = "<role>"` field (when Codex compatibility is required). ## Steps 1. Start from the current `.github/instructions/project.instructions.md` when it already exists and passed the official-init success bar: - Reuse that file as the primary fact packet. - Scan only the missing fact categories, missing scaffold surfaces, or unresolved `unknown` gaps. 2. Otherwise do a bounded manual scan: - Read `CLAUDE.md`, `.github/copilot-instructions.md`, or `AGENTS.md` first when `repo-init-official` created or updated one. - Read `README*`, package/build files, CI files, app entrypoints, test directories. - Record unresolved facts as `unknown` instead of guessing. 3. **PM must analyze the project** to produce real project facts in `.github/instructions/project.instructions.md`. Skip this step only if the file already came from official init (step 1) and has no more than 3 "unknown" fields. Otherwise: - If the file already exists from the shell helper, read it and enhance any "unknown" fields. - If the file does not exist, analyze the project first, then create it. - Read `pom.xml`/`build.gradle`/`package.json` for build system, modules, dependencies. - For Maven: parse `<modules>`, find main module, find `@SpringBootApplication`. - Search for security/auth files, runtime config (`application.yml`, Docker files). - Infer build/test commands. Only mark genuinely undetectable items as `unknown`. 4. Create or repair `.github/instructions/project.instructions.md` with: `Project Facts`, `Build and Test Commands`, `Runtime and Entry Points`, `Module Boundaries`, `Known Unknowns`, `Verification Notes`, and `Init Status`. 5. Replace markers with `Init source: manual_fallback|official_init_plus_manual_fallback` and a real timestamp. 6. Ensure `## Init Status` contains: `Init ready`, `Required artifacts verified`, `Bootstrap contract version: 0.8.0`, `Last full verification`, `Reentry rule`. 7. Initialize missing scaffolds: shell helper → `target-instructions-bootstrap` → `target-memory-bootstrap` → `target-spec-bootstrap`. 8. Re-check the required artifact paths on disk. 9. Re-check the required instruction content. 10. After all other artifacts and content checks pass, write `best-copilot.md` with the exact sentinel. 11. Verify: `project.instructions.md` exists and is not placeholder-heavy, `Init Status` has version `0.8.0`, `best-copilot.md` matches sentinel, and `verified_paths` covers the full runtime-required path set. ## Inline Fallback Templates When the shell helper and bootstrap skills are unavailable, read `inline-templates-reference.md` in this skill directory for the canonical templates. Do NOT improvise content. ## Output ```markdown ## Repo Init Manual Fallback - status: success|blocked - required_artifacts_verified: yes|no - sentinel_written: yes|no - next_task_ready: yes|no - verified_paths: <every required artifact checked on disk, plus best-copilot.md> - missing_paths: none|<paths not present or not content-valid> ```
From funky-eyes/best-copilot, Apache-2.0.
More from funky-eyes/best-copilot
- 0.599score
evolution-loop
funky-eyes
Use after meaningful task closeout, repeated failure, review loops, stale triggers, or user feedback that an agent/skill/workflow should improve.
SkillBFreshness: B★ 12Claude Code +9 - 0.599score
target-instructions-bootstrap
funky-eyes
Create the target repository's local AI instruction scaffold during first-use bootstrap.
SkillBFreshness: B★ 12Claude Code +9 - 0.599score
td-python-coding-guidelines
funky-eyes
Use when writing or reviewing Python code, fixing lint/docstring/import/naming/exception/type annotation issues, unifying team style, or migrating `.py` files…
SkillBFreshness: B★ 12Claude Code +9 - 0.599score
structured-review
funky-eyes
Use when reviewing code diffs, implementation changes, agent/skill/instruction customization, targeted re-review, review feedback intake, review handoff…
SkillBFreshness: B★ 12Claude Code +9 - 0.599score
td-java-coding-guidelines
funky-eyes
Use when writing or reviewing Java code involving Tongdun/Alibaba Java rules, exceptions, logging, SQL/MyBatis, security, middleware, concurrency, virtual…
SkillBFreshness: B★ 12Claude Code +9 - 0.599score
core-workflow-contract
funky-eyes
Use when an agent needs the shared best-copilot contract for source priority, runtime adapters, init gates, work modes, dispatch packets, review…
SkillBFreshness: B★ 12Claude Code +9 - 0.596score
target-memory-bootstrap
funky-eyes
Create the target repository's local memory skeleton after repo init.
SkillBFreshness: B★ 12Claude Code +9 - 0.596score
repo-init-official
funky-eyes
Use inside repo-init-scan to try a target-local `init` skill, Claude `/init`, or `copilot init`, normalize official init output into…
SkillBFreshness: B★ 12Claude Code +9 - 0.596score
repo-init-gate
funky-eyes
Use before repo-init-scan to cheaply decide whether a target repository already satisfies the current best-copilot init contract by reading only the root…
SkillBFreshness: B★ 12Claude Code +9 - 0.596score
target-spec-bootstrap
funky-eyes
Create the target repository's local spec skeleton and reusable requirements/design/tasks templates.
SkillBFreshness: B★ 12Claude Code +9 - 0.594score
gitnexus-guide
funky-eyes
Use when the user asks about GitNexus itself — available tools, how to query the knowledge graph, MCP resources, graph schema, or workflow reference.
SkillBFreshness: B★ 12Claude Code +9 - 0.594score
security-reviewer-workflow
funky-eyes
Use when reviewing authentication, authorization, dependencies, configuration, release surfaces, sensitive data, logging, validation, CORS, secrets, or…
SkillBFreshness: B★ 12Claude Code +9
Signals
- Stars
- 12
- Forks
- 1
- Open issues
- 0
- Last commit (this folder)
- Aug 4, 2026 (2 months ago)
- Last release
- Aug 3, 2026 (2 months ago)
From GitHub (fetched Oct 24, 2026).
Details
- License
- Apache-2.0
- Repository
- github.com/funky-eyes/best-copilot
- First seen
- Oct 24, 2026
- Last checked
- Oct 24, 2026
- Source
- GitHubgithub-graphql
More from funky-eyes
- 0.599score
evolution-loop
funky-eyes
Use after meaningful task closeout, repeated failure, review loops, stale triggers, or user feedback that an agent/skill/workflow should improve.
SkillBFreshness: B★ 12Claude Code +9 - 0.599score
td-java-coding-guidelines
funky-eyes
Use when writing or reviewing Java code involving Tongdun/Alibaba Java rules, exceptions, logging, SQL/MyBatis, security, middleware, concurrency, virtual…
SkillBFreshness: B★ 12Claude Code +9 - 0.599score
core-workflow-contract
funky-eyes
Use when an agent needs the shared best-copilot contract for source priority, runtime adapters, init gates, work modes, dispatch packets, review…
SkillBFreshness: B★ 12Claude Code +9 - 0.599score
td-python-coding-guidelines
funky-eyes
Use when writing or reviewing Python code, fixing lint/docstring/import/naming/exception/type annotation issues, unifying team style, or migrating `.py` files…
SkillBFreshness: B★ 12Claude Code +9 - 0.599score
target-instructions-bootstrap
funky-eyes
Create the target repository's local AI instruction scaffold during first-use bootstrap.
SkillBFreshness: B★ 12Claude Code +9 - 0.599score
structured-review
funky-eyes
Use when reviewing code diffs, implementation changes, agent/skill/instruction customization, targeted re-review, review feedback intake, review handoff…
SkillBFreshness: B★ 12Claude Code +9
Done for you
Rather not set it up yourself?
Chalk Labs scopes, builds and runs agents for your team: the task audit, the eval harness, the integrations and the weekly ops report. Fixed scope, priced before we start.
See how we build AI agents →Read the case studies
Chalk Labs is a two-founder AI and growth studio. We build AI agents, CRMs and MVPs for Web3 and AI companies, and have shipped 50+ product and token launches.
- launches shipped
- 50+
- launches shipped
- media placements
- 500+
- media placements
- founders on every account
- 2
- founders on every account
Press we've placed: Forbes · CoinDesk · Decrypt · Cointelegraph · The Block
Rahil Jain, Co-founder, Chalk Labs. Ex-Goldman Sachs engineer. You talk to the person who builds it.
Get repo-init-manual-fallback running in your business
Tell us what you want to automate. A founder replies within one working day with a plan, not a pitch deck.
Cite this page
repo-init-manual-fallback skill. Agent Mart. https://agentmart.dev/skills/funky-eyes-repo-init-manual-fallback (last updated 2026-10-24)