[{"data":1,"prerenderedAt":-1},["ShallowReactive",2],{"skill:01M7ERCV5D7B1MN53AXBVVS9NE":3,"skill-same:01M7ERCV5D7B1MN53AXBVVS9NE":48},{"skillNo":4,"name":5,"nameZh":6,"description":7,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":16,"categories":17,"keywords":18,"publishedAt":19,"linkOnly":20,"markdownContent":21,"markdownBody":22,"frontMatter":23,"refName":26,"commitSha":27,"skillDirectory":28,"forksCount":29,"repositoryHtmlUrl":30,"repositoryHomepageUrl":6,"ownerHtmlUrl":31,"ownerBlogUrl":6,"ownerDescription":32,"updatedAt":33,"install":34,"resources":47},"01M7ERCV5D7B1MN53AXBVVS9NE","maintain-formal-verification-skill",null,"Bring a project's formal verification back in line after the code changed: update the model, rerun, and say whether each failure is drift or a bug. Use for 'is the model still true to the code', or after merges touched verified code.","JoshuaHayesWasHere\u002Fjstack","JoshuaHayesWasHere","https:\u002F\u002Favatars.githubusercontent.com\u002Fu\u002F78288453?v=4","01M7EQPZP0GDYN3SYY9SZZHFJE","Joshua Hayes","01M7EQPZPQKJDZR99NG512A8X2",1,"MIT","https:\u002F\u002Fgithub.com\u002FJoshuaHayesWasHere\u002Fjstack\u002Fblob\u002Fmain\u002Ffeatures\u002Fskill-library\u002Fskills\u002Fmaintain-formal-verification-skill\u002FSKILL.md",[],[],"2026-11-15 14:19:17",false,"---\nname: maintain-formal-verification-skill\ndescription: \"Bring a project's formal verification back in line after the code changed: update the model, rerun, and say whether each failure is drift or a bug. Use for 'is the model still true to the code', or after merges touched verified code.\"\ntip: brings a formal verify skill back in line with the code after changes that skipped it | is the model still true to the code | maintain the formal verification | formal verification drift | recheck the model\ncard: \"Distrust the change | 8 | Brings a formal verify skill back in line with the code: reads every change to covered code since the sync commit, updates the model or harness from the code, reruns, and says of each failure whether it is drift or a bug. | Reach for it after merges or refactors touched verified code without rechecking it. | op: the operator\"\n---\n\n# Maintain a formal verification skill\n\nSource: pstack, `pstack\u002Fskills\u002Fmaintain-formal-verification-skill\u002FSKILL.md`\n\nA formal check proves a property of what it describes.\nWhen the code changes and the verification does not, the check can keep passing while it describes code that no longer exists.\nThat pass is the dangerous case, because it looks like a proof.\nThis skill is the upkeep loop for a verify skill that `\u002Fcreate-formal-verification-skill` generated, or for any project-local skill that runs a model checker, prover, or code-level verifier against named code.\nThe project-local skill handles a change that an agent makes while it uses that skill.\nThis skill handles the changes that skipped it: merges, edits made without the skill, and drift that builds up across many commits.\n\n## Where it runs\n\nRun this in a durable thread, never in a subagent, whose context ends with its answer: the reruns and the breaks that prove each updated check can still fail take many turns, and the updated verify skill is the part that has to outlive them.\nA crewmate under a brief does all of it, in its own worktree, and every file this skill changes lands there.\nA scout under a brief may do steps 1 to 3, read-only, and report each check's outcome so far and the changes that need an update, since a scout changes nothing in the repo: the crewmate that follows updates the verification and reruns it.\nThe Result (or the scout's Report) carries what the report below asks for.\n\n## Outcomes\n\nGive each check one outcome, and say which:\n\n- **in sync**: no change since the last sync alters what the check describes, and the check passes.\n- **updated**: the code changed, the verification now describes the new code, and the check passes.\n- **bug**: the verification describes the current code, and the code breaks the property.\n  This is a finding about the code.\n  It is not drift.\n- **blocked**: the check could not be brought back in line or could not run.\n  Say exactly why.\n\nA check can be both updated and bug: the update makes the verification describe the code again, and then the property fails on it.\n\n## Edit scope\n\nEdit only the verification: models, harnesses, property files, and the verify skill itself.\nNever edit product code during a run, unless the user asks you to.\nWhen a property fails on the current code, do not weaken the property or the model to make it pass.\nReport the failure.\n\n## Pass\n\n1. **Find the checks.**\n   A verify skill has no fixed name, so find it by what it does: a project skill that runs a model checker, a prover, or another verifier against named code.\n   In a repo that keeps its project skills in its features, as jstack does, each is an item of `skills.json` under `projectSkills`, whose `path` is the skill's folder; otherwise look in the repo's own project-skill folder, such as `.claude\u002Fskills\u002F`.\n   Read each one you find, or the one the user named.\n   For each check, collect the property, the code it covers, the bounds and assumptions, the command that runs it, and the commit where it was last confirmed to describe the code.\n   If no verify skill exists, stop and point at `\u002Fcreate-formal-verification-skill`.\n   If the skill does not record a sync commit, use the last commit that changed the verification files, and say that you did.\n\n2. **Run it as it is.**\n   Run each check once before you change anything, and record the result.\n   A pass here does not tell you that the code is correct.\n   It tells you that the old description still holds.\n   A failure here can be drift or a bug.\n   Do not decide which until step 5.\n\n3. **Read every change to covered code.**\n   Diff the covered files from the sync commit to the current commit, and read each commit message.\n   Also look for code that moved: a covered function that is renamed, split, or moved to another file is still covered code.\n   Also look for new code that touches the state that a check guards, such as a new function that writes the same fields.\n   For each change, decide which of these it is:\n   - **No effect.**\n     The change does not alter anything the verification describes, such as formatting, comments, or code paths that the verification does not cover.\n     Check this against the model or harness, not against the commit message.\n   - **The verification no longer describes the code.**\n     A step that was one indivisible action is now two, a branch was added or removed, a check moved, a new path reaches the guarded state, or a value now has a different range.\n     The model or harness must change.\n   - **The promise changed on purpose.**\n     The change alters what the code promises, and evidence in the change says that this is intended: the doc comment, the tests, the changelog, or the commit message.\n     The property must change, and you must cite that evidence.\n   - **The promise may be broken.**\n     The code changed, and no evidence says that the promise changed.\n     Treat a failure that follows as a bug candidate.\n\n4. **Update from the code, not from the result.**\n   Change the model or harness so that each changed part again points to code that a reviewer can open, and keep one indivisible step in the code as one step in the verification.\n   Write the update from the code before you rerun, and do not shape it by what makes the check pass.\n   Change a property only for a promise that changed on purpose.\n   Write down the old wording, the new wording, and the evidence.\n   Keep the bounds unless the change makes them wrong.\n   If you must reduce a bound, say so in the report.\n\n5. **Rerun and judge each result.**\n   - **Pass:** the check can now pass on a description that has lost the ability to fail.\n     Break the new code the way the property guards against, such as removing a new check or reordering a new step, and confirm that the check catches it.\n     Make the break in this thread's own worktree, never in the main checkout or in a worktree that another thread uses, and leave it uncommitted.\n     Then drop the break (`git restore` the files it touched).\n   - **Fail:** replay the counterexample against the real code, step by step.\n     If a step cannot happen in the code, the verification is wrong.\n     Fix the verification, and say what was wrong.\n     If every step can happen, it is a bug.\n     Name the commit that introduced it, tell the counterexample as a sequence of steps in the code's own words, and state the harm in one sentence.\n     If a cheap test or reproduction against the real code shows the bug, add it to the report as evidence.\n     Do not commit it to the product code.\n\n6. **Update the verify skill.**\n   Update the code that each check covers, the commands, the bounds, the assumptions, and the sync commit.\n   Set the sync commit to the commit that you checked.\n   Do not move the sync commit for a blocked check, and say so in the report.\n   Then follow the verify skill once as the next agent would, so that you know its steps still run.\n\n## Report\n\nTell the user, for each check:\n\n- the outcome, and the commits that you read;\n- what changed in the verification and why, with a short reason for each change: the code changed shape, or the promise changed on purpose (with the evidence);\n- the rerun result, and the break that you used to show the updated check can still fail;\n- each bug, in plain words: the counterexample as steps in the code's words, the harm, and the commit that introduced it;\n- what the verification still does not cover, including new code that it should cover and does not.\n\nSay outright which failures were drift in the verification and which were bugs in the code.\nFor a check that passed before you updated it and fails after, say that the old pass was on a stale description.\n\nPut the verification updates in a change of their own, separate from any product fix.\n","# Maintain a formal verification skill\n\nSource: pstack, `pstack\u002Fskills\u002Fmaintain-formal-verification-skill\u002FSKILL.md`\n\nA formal check proves a property of what it describes.\nWhen the code changes and the verification does not, the check can keep passing while it describes code that no longer exists.\nThat pass is the dangerous case, because it looks like a proof.\nThis skill is the upkeep loop for a verify skill that `\u002Fcreate-formal-verification-skill` generated, or for any project-local skill that runs a model checker, prover, or code-level verifier against named code.\nThe project-local skill handles a change that an agent makes while it uses that skill.\nThis skill handles the changes that skipped it: merges, edits made without the skill, and drift that builds up across many commits.\n\n## Where it runs\n\nRun this in a durable thread, never in a subagent, whose context ends with its answer: the reruns and the breaks that prove each updated check can still fail take many turns, and the updated verify skill is the part that has to outlive them.\nA crewmate under a brief does all of it, in its own worktree, and every file this skill changes lands there.\nA scout under a brief may do steps 1 to 3, read-only, and report each check's outcome so far and the changes that need an update, since a scout changes nothing in the repo: the crewmate that follows updates the verification and reruns it.\nThe Result (or the scout's Report) carries what the report below asks for.\n\n## Outcomes\n\nGive each check one outcome, and say which:\n\n- **in sync**: no change since the last sync alters what the check describes, and the check passes.\n- **updated**: the code changed, the verification now describes the new code, and the check passes.\n- **bug**: the verification describes the current code, and the code breaks the property.\n  This is a finding about the code.\n  It is not drift.\n- **blocked**: the check could not be brought back in line or could not run.\n  Say exactly why.\n\nA check can be both updated and bug: the update makes the verification describe the code again, and then the property fails on it.\n\n## Edit scope\n\nEdit only the verification: models, harnesses, property files, and the verify skill itself.\nNever edit product code during a run, unless the user asks you to.\nWhen a property fails on the current code, do not weaken the property or the model to make it pass.\nReport the failure.\n\n## Pass\n\n1. **Find the checks.**\n   A verify skill has no fixed name, so find it by what it does: a project skill that runs a model checker, a prover, or another verifier against named code.\n   In a repo that keeps its project skills in its features, as jstack does, each is an item of `skills.json` under `projectSkills`, whose `path` is the skill's folder; otherwise look in the repo's own project-skill folder, such as `.claude\u002Fskills\u002F`.\n   Read each one you find, or the one the user named.\n   For each check, collect the property, the code it covers, the bounds and assumptions, the command that runs it, and the commit where it was last confirmed to describe the code.\n   If no verify skill exists, stop and point at `\u002Fcreate-formal-verification-skill`.\n   If the skill does not record a sync commit, use the last commit that changed the verification files, and say that you did.\n\n2. **Run it as it is.**\n   Run each check once before you change anything, and record the result.\n   A pass here does not tell you that the code is correct.\n   It tells you that the old description still holds.\n   A failure here can be drift or a bug.\n   Do not decide which until step 5.\n\n3. **Read every change to covered code.**\n   Diff the covered files from the sync commit to the current commit, and read each commit message.\n   Also look for code that moved: a covered function that is renamed, split, or moved to another file is still covered code.\n   Also look for new code that touches the state that a check guards, such as a new function that writes the same fields.\n   For each change, decide which of these it is:\n   - **No effect.**\n     The change does not alter anything the verification describes, such as formatting, comments, or code paths that the verification does not cover.\n     Check this against the model or harness, not against the commit message.\n   - **The verification no longer describes the code.**\n     A step that was one indivisible action is now two, a branch was added or removed, a check moved, a new path reaches the guarded state, or a value now has a different range.\n     The model or harness must change.\n   - **The promise changed on purpose.**\n     The change alters what the code promises, and evidence in the change says that this is intended: the doc comment, the tests, the changelog, or the commit message.\n     The property must change, and you must cite that evidence.\n   - **The promise may be broken.**\n     The code changed, and no evidence says that the promise changed.\n     Treat a failure that follows as a bug candidate.\n\n4. **Update from the code, not from the result.**\n   Change the model or harness so that each changed part again points to code that a reviewer can open, and keep one indivisible step in the code as one step in the verification.\n   Write the update from the code before you rerun, and do not shape it by what makes the check pass.\n   Change a property only for a promise that changed on purpose.\n   Write down the old wording, the new wording, and the evidence.\n   Keep the bounds unless the change makes them wrong.\n   If you must reduce a bound, say so in the report.\n\n5. **Rerun and judge each result.**\n   - **Pass:** the check can now pass on a description that has lost the ability to fail.\n     Break the new code the way the property guards against, such as removing a new check or reordering a new step, and confirm that the check catches it.\n     Make the break in this thread's own worktree, never in the main checkout or in a worktree that another thread uses, and leave it uncommitted.\n     Then drop the break (`git restore` the files it touched).\n   - **Fail:** replay the counterexample against the real code, step by step.\n     If a step cannot happen in the code, the verification is wrong.\n     Fix the verification, and say what was wrong.\n     If every step can happen, it is a bug.\n     Name the commit that introduced it, tell the counterexample as a sequence of steps in the code's own words, and state the harm in one sentence.\n     If a cheap test or reproduction against the real code shows the bug, add it to the report as evidence.\n     Do not commit it to the product code.\n\n6. **Update the verify skill.**\n   Update the code that each check covers, the commands, the bounds, the assumptions, and the sync commit.\n   Set the sync commit to the commit that you checked.\n   Do not move the sync commit for a blocked check, and say so in the report.\n   Then follow the verify skill once as the next agent would, so that you know its steps still run.\n\n## Report\n\nTell the user, for each check:\n\n- the outcome, and the commits that you read;\n- what changed in the verification and why, with a short reason for each change: the code changed shape, or the promise changed on purpose (with the evidence);\n- the rerun result, and the break that you used to show the updated check can still fail;\n- each bug, in plain words: the counterexample as steps in the code's words, the harm, and the commit that introduced it;\n- what the verification still does not cover, including new code that it should cover and does not.\n\nSay outright which failures were drift in the verification and which were bugs in the code.\nFor a check that passed before you updated it and fails after, say that the old pass was on a stale description.\n\nPut the verification updates in a change of their own, separate from any product fix.\n",{"tip":24,"card":25,"name":5,"description":7},"brings a formal verify skill back in line with the code after changes that skipped it | is the model still true to the code | maintain the formal verification | formal verification drift | recheck the model","Distrust the change | 8 | Brings a formal verify skill back in line with the code: reads every change to covered code since the sync commit, updates the model or harness from the code, reruns, and says of each failure whether it is drift or a bug. | Reach for it after merges or refactors touched verified code without rechecking it. | op: the operator","main","daff051d2674b46e2eb8a333bbd3f3e40f8d1cea","features\u002Fskill-library\u002Fskills\u002Fmaintain-formal-verification-skill",0,"https:\u002F\u002Fgithub.com\u002FJoshuaHayesWasHere\u002Fjstack","https:\u002F\u002Fgithub.com\u002FJoshuaHayesWasHere","Software Engineer @ Markel\r\n","2026-11-15 11:25:06",{"source":8,"skillName":5,"commands":35,"prompt":45,"zipAvailable":46},[36,39,42],{"tool":37,"command":38},"npx","npx skills add JoshuaHayesWasHere\u002Fjstack --skill maintain-formal-verification-skill",{"tool":40,"command":41},"bunx","bunx skills add JoshuaHayesWasHere\u002Fjstack --skill maintain-formal-verification-skill",{"tool":43,"command":44},"pnpm","pnpm dlx skills add JoshuaHayesWasHere\u002Fjstack --skill maintain-formal-verification-skill","安装技能，技能地址：https:\u002F\u002Fgithub.com\u002FJoshuaHayesWasHere\u002Fjstack\u002Fblob\u002Fmain\u002Ffeatures\u002Fskill-library\u002Fskills\u002Fmaintain-formal-verification-skill\u002FSKILL.md",true,[],[49,56,63,70,77,84],{"skillNo":50,"name":51,"nameZh":6,"description":52,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":53,"categories":54,"keywords":55,"publishedAt":19},"01M7ERF9GXMY62V7QTVAR6G8DG","no-key","A skill its skills.json lists as hidden that a model may still call by itself, kept broken so the lint is seen to catch it.","https:\u002F\u002Fgithub.com\u002FJoshuaHayesWasHere\u002Fjstack\u002Fblob\u002Fmain\u002Ftesting\u002Ftests\u002Ffixtures\u002Fskills\u002Fno-key\u002FSKILL.md",[],[],{"skillNo":57,"name":58,"nameZh":6,"description":59,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":60,"categories":61,"keywords":62,"publishedAt":19},"01M7ERF83ZG3XKJ2SB8MJHCZ8H","fake-skill","A FAKE skill in a FAKE tree, named for its folder until the test renames the folder.","https:\u002F\u002Fgithub.com\u002FJoshuaHayesWasHere\u002Fjstack\u002Fblob\u002Fmain\u002Ftesting\u002Ftests\u002Ffixtures\u002Fskills\u002Fnamed-tree\u002Fskills\u002Ffake-skill\u002FSKILL.md",[],[],{"skillNo":64,"name":65,"nameZh":6,"description":66,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":67,"categories":68,"keywords":69,"publishedAt":19},"01M7ERF698RV4VF8QCDF8H8DA4","extra-key","A skill its skills.json does not list as hidden that a model still cannot call, kept broken so the lint is seen to catch it.","https:\u002F\u002Fgithub.com\u002FJoshuaHayesWasHere\u002Fjstack\u002Fblob\u002Fmain\u002Ftesting\u002Ftests\u002Ffixtures\u002Fskills\u002Fextra-key\u002FSKILL.md",[],[],{"skillNo":71,"name":72,"nameZh":6,"description":73,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":74,"categories":75,"keywords":76,"publishedAt":19},"01M7ERF2V8N73X9T0Y9PJ4SFV6","unslop","Cut AI tells from writing. Use when the user says text reads like AI, asks to unslop it, or before prose ships to people.","https:\u002F\u002Fgithub.com\u002FJoshuaHayesWasHere\u002Fjstack\u002Fblob\u002Fmain\u002Ffeatures\u002Fwriting\u002Fskills\u002Funslop\u002FSKILL.md",[],[],{"skillNo":78,"name":79,"nameZh":6,"description":80,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":81,"categories":82,"keywords":83,"publishedAt":19},"01M7EREWM67WX934XKABM4XCSZ","technical-writing","Write or review docs, READMEs, RFCs, PR descriptions and commit messages to a layered standard: Diataxis structure, Google developer style, STE rules and Global English.","https:\u002F\u002Fgithub.com\u002FJoshuaHayesWasHere\u002Fjstack\u002Fblob\u002Fmain\u002Ffeatures\u002Fwriting\u002Fskills\u002Ftechnical-writing\u002FSKILL.md",[],[],{"skillNo":85,"name":86,"nameZh":6,"description":87,"descriptionZh":6,"repositoryFullName":8,"ownerLogin":9,"ownerAvatarUrl":10,"ownerAccountNo":11,"ownerDisplayName":12,"repositoryNo":13,"starsCount":14,"licenseSpdxId":15,"licenseFileUrl":6,"sourceHtmlUrl":88,"categories":89,"keywords":90,"publishedAt":19},"01M7EREMHJS9JY10AM5EKVX3GP","why","Find why code is the way it is, from its history and every evidence source this session can query, with citations. Use for 'why does X work this way', 'why did we pick Y', regressions and postmortems. Use how for runtime behavior.","https:\u002F\u002Fgithub.com\u002FJoshuaHayesWasHere\u002Fjstack\u002Fblob\u002Fmain\u002Ffeatures\u002Fwhy\u002FSKILL.md",[],[]]