setup-promise-to-proof
Skill4 starsChanged 36 days agoApache-2.0View on GitHub
Set up ignored .p2p workflow storage, with optional tracker, triage labels, and domain-doc pointers.
In this repository19 other files it gives its agents.
- AGENTS.mdAGENTS.md
- audit-acceptanceSkill
- create-parent-issueSkill
- critiqueSkill
- deliver-issueSkill
- fix-prSkill
- implement-contractSkill
- interrogateSkill
- merge-readinessSkill
- plan-acceptanceSkill
- proveSkill
- publish-pr-reviewSkill
- publish-prSkill
- repair-gapsSkill
- replan-backlogSkill
- retrospectSkill
- review-implementationSkill
- slice-contractSkill
- triage-issueSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 1 is distinctive, and 5 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
5
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 17
Read existing AGENTS.md or CLAUDE.md, docs/agents/, Git remotes, and any existing issue or label conventions.
- Very specific: line 50
Create .p2p/work/ and .p2p/tmp/ without replacing existing content.
- Very specific: line 53
Check effective ignore rules for probes under both .p2p directories, including repository, global, and .git/info/exclude rules.
- Very specific: line 61
Use p2p-state/<slug>.json as the default portable checkpoint destination.
- Very specific: line 69
Create or update resolved docs/agents/ configuration files.
- Distinctive: line 76
Read back each changed file and check the pointers, destination, and all five label mappings.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Set up Promise to Proof
Configure the repository where the skills will run, not the skill collection's source repository. Invocation authorizes the standard local filesystem setup. Inspect and preserve existing files; ask only about unresolved consequential configuration choices. Do not create tracker issues or labels, change remotes, or create domain documents during setup.
Inspect the repository
Read existing AGENTS.md or CLAUDE.md, docs/agents/, Git remotes, and any
existing issue or label conventions. Look for CONTEXT.md, CONTEXT-MAP.md,
and relevant ADR locations. When possible, inspect the destination's existing
GitHub labels with gh before proposing a mapping. Keep existing configuration
and human edits; a rerun must not replace them with defaults.
Choose the configuration
- Issue tracker (optional): Local planning and delivery need no tracker.
Configure GitHub only when requested or already in use. Confirm the
target GitHub repository from the remote or ask for it when ambiguous. Draft
docs/agents/issue-tracker.mdwith the destination, how to read, search, create, edit, label, and verify issues usinggh, and whether PRs count as incoming requests (default: no). If the project uses another tracker, report that its tracker-writing skills are not ready; do not claim that a generic tracker configuration makes the GitHub-specific skills work. - Triage labels (only with a tracker): Draft
docs/agents/triage-labels.mdmapping the five rolesneeds-triage,needs-info,ready-for-agent,ready-for-human, andwontfixto actual destination label strings. Recommend identical names when the project has no mapping; confirm overrides before writing. A file alone does not create labels in GitHub: report missing labels separately. - Domain docs: Draft
docs/agents/domain.mddescribing where to find the project's glossary/context and architectural decisions. Use the existing layout, or propose rootCONTEXT.mdanddocs/adr/for a single-context project. Treat absent domain files as optional; do not create placeholderCONTEXT.md,CONTEXT-MAP.md, or ADRs.
Resolve repository conventions from existing configuration. Present any unresolved tracker destination, label overrides, or domain layout choices for a decision; continue independent local filesystem setup.
Write and verify
Create .p2p/work/ and .p2p/tmp/ without replacing existing content. Do
not create specs/ or work/; projects may use either directory for their own
files. Add /.p2p/ to .gitignore, preserving unrelated rules and comments.
Check effective ignore rules for probes under both .p2p directories, including
repository, global, and .git/info/exclude rules. The repository rule must
ignore .p2p state.
Report setup errors and use git check-ignore -v to locate conflicting patterns;
do not hide them with force-add or silently rewrite unrelated policy. Empty directories need no
placeholder files. Run python3 <skill-dir>/scripts/p2p_filesystem.py --repo <root> setup;
its error output identifies conflicts.
Use p2p-state/<slug>.json as the default portable checkpoint destination;
keep it trackable and reserve it for compact P2P metadata. Do not create empty
checkpoint files. An explicitly selected GitHub issue is the alternative per
work item; do not establish competing synchronized stores. Explain that local
writes are not remote preservation, and that candidate/source commits must also
be retrievable before portability is claimed. Neither setup nor checkpoint
creation authorizes commits, pushes or issue comments.
Create or update resolved docs/agents/ configuration files. Add
short pointers to them under ## Agent skills in an existing AGENTS.md or
CLAUDE.md; if both exist, confirm which is authoritative. If neither exists,
ask whether to create AGENTS.md. Preserve other instructions and update
existing pointers in place instead of appending duplicates. Point only to
files actually written.
Read back each changed file and check the pointers, destination, and all five label mappings. Report what was configured, any missing labels or unsupported tracker, and what still needs a decision. A repeat invocation with unchanged answers should leave the files unchanged.
End with Next steps: and a numbered list (1., 2., ...) of applicable
actions in order, so each can be referenced by number. If approval is pending,
name the exact files and decision. For each missing GitHub label, give
gh label create "<label>" --repo <owner/repo>. For an unsupported tracker,
name that tracker and the setup capability it lacks. Once setup is verified, give /plan-acceptance <project source path or agreed text>;
optional source files may live in specs/, work/, or another project path.
Generated contracts go under .p2p/work/<slug>/contract.md. Optionally give
/triage-issue <issue URL> when the user has chosen a tracker. Saving setup files never stages, commits,
pushes, or creates external records.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

